{"topics":[{"id":"designing-and-planning","name":"Designing and Planning","set":"core","count":22},{"id":"provisioning-and-infrastructure","name":"Provisioning and Infrastructure","set":"core","count":22},{"id":"security-and-compliance","name":"Security and Compliance","set":"core","count":22},{"id":"implementation-and-migration","name":"Implementation and Migration","set":"core","count":22},{"id":"optimization-and-cost","name":"Optimization and Cost","set":"core","count":22},{"id":"reliability-and-operations","name":"Reliability and Operations","set":"core","count":22},{"id":"case-study-altostrat-media","name":"Case study: Altostrat Media","set":"core","count":22},{"id":"case-study-cymbal-retail","name":"Case study: Cymbal Retail","set":"core","count":22},{"id":"case-study-ehr-healthcare","name":"Case study: EHR Healthcare","set":"core","count":22},{"id":"case-study-knightmotives-automotive","name":"Case study: KnightMotives Automotive","set":"core","count":22},{"id":"ai-ml-and-generative-ai","name":"AI/ML and Generative AI","set":"advanced","count":10},{"id":"vpc-service-controls-and-advanced-security","name":"VPC Service Controls & Advanced Security","set":"advanced","count":10},{"id":"anthos-gke-enterprise-multi-cloud","name":"Anthos / GKE Enterprise / Multi-cloud","set":"advanced","count":8},{"id":"advanced-networking","name":"Advanced Networking","set":"advanced","count":8},{"id":"advanced-database-comparison","name":"Advanced Database Comparison","set":"advanced","count":8},{"id":"advanced-cost-optimization","name":"Advanced Cost Optimization","set":"advanced","count":8},{"id":"advanced-reliability-and-operations","name":"Advanced Reliability & Operations","set":"advanced","count":8},{"id":"hybrid-multi-cloud-and-migration","name":"Hybrid/Multi-cloud & Migration","set":"advanced","count":8},{"id":"bigquery-advanced-and-data-analytics","name":"BigQuery Advanced & Data Analytics","set":"advanced","count":6},{"id":"well-architected-framework-and-architecture-patterns","name":"Well-Architected Framework & Architecture Patterns","set":"advanced","count":6}],"questions":[{"id":"1","label":"Q1","set":"core","topicId":"designing-and-planning","text":"A startup needs a managed relational database for a global customer portal. It requires horizontal scale, strong consistency, and 99.999% availability. Which service should you choose?","options":["Cloud SQL for MySQL","Cloud Spanner","Bigtable","Firestore"],"answer":[1],"why":"Cloud Spanner is the managed relational database designed for horizontal scale, strong consistency, and very high availability."},{"id":"2","label":"Q2","set":"core","topicId":"designing-and-planning","text":"A media company wants to run stateless web services with automatic scaling and minimal ops. Requests are HTTP-based and startup time is short. What is the best compute platform?","options":["Compute Engine managed instance group","GKE Standard","Cloud Run","Bare metal solution"],"answer":[2],"why":"Cloud Run is best for stateless HTTP services when you want autoscaling with minimal operational overhead."},{"id":"3","label":"Q3","set":"core","topicId":"designing-and-planning","text":"A team wants a serverless event-driven architecture where messages must be decoupled from producers and consumers. Which service best fits?","options":["Pub/Sub","Cloud Tasks","Memorystore","Cloud NAT"],"answer":[0],"why":"Pub/Sub is the managed messaging service for asynchronous decoupling in event-driven systems."},{"id":"4","label":"Q4","set":"core","topicId":"designing-and-planning","text":"A company wants to store time-series sensor data with very high write throughput and low-latency key-based reads. Which database is best?","options":["Bigtable","Cloud SQL","Firestore in Datastore mode","Cloud Storage"],"answer":[0],"why":"Bigtable is optimized for very high throughput, sparse wide tables, and time-series or IoT-style data."},{"id":"5","label":"Q5","set":"core","topicId":"designing-and-planning","text":"A company needs an object store for unstructured documents, backups, and images with lifecycle policies. Which service should you choose?","options":["Persistent Disk","Cloud Storage","Filestore","BigQuery"],"answer":[1],"why":"Cloud Storage is the managed object storage service with lifecycle management and multiple storage classes."},{"id":"6","label":"Q6","set":"core","topicId":"designing-and-planning","text":"A business wants to analyze petabytes of structured and semi-structured data using SQL without managing infrastructure. Which service is best?","options":["BigQuery","Cloud SQL","Bigtable","Dataproc"],"answer":[0],"why":"BigQuery is the serverless analytics warehouse for large-scale SQL analytics."},{"id":"7","label":"Q7","set":"core","topicId":"designing-and-planning","text":"A company is migrating VMs from on-premises with the least redesign first. Which approach is best?","options":["Rewrite all apps to Cloud Run","Lift and shift to Compute Engine","Move directly to GKE Autopilot","Rebuild on Cloud Functions"],"answer":[1],"why":"When redesign is not yet feasible, lift-and-shift to Compute Engine is usually the lowest-risk first step."},{"id":"8","label":"Q8","set":"core","topicId":"designing-and-planning","text":"A team needs shared POSIX file storage for legacy applications running on Compute Engine. Which service should they use?","options":["Cloud Storage","Filestore","Persistent Disk snapshots","Pub/Sub"],"answer":[1],"why":"Filestore provides managed NFS for applications that require shared POSIX-compatible file storage."},{"id":"9","label":"Q9","set":"core","topicId":"designing-and-planning","text":"A company must isolate workloads across multiple environments while centralizing billing and governance. What should they use?","options":["Separate projects under folders and an organization","Single project with labels","One VPC per team only","Multiple billing accounts only"],"answer":[0],"why":"Projects provide isolation, folders support hierarchy, and the organization enables centralized governance."},{"id":"10","label":"Q10","set":"core","topicId":"designing-and-planning","text":"A banking app needs global HTTP load balancing with SSL termination and CDN integration. Which solution is best?","options":["Regional internal TCP/UDP load balancer","External Application Load Balancer","Proxyless gRPC only","Cloud NAT"],"answer":[1],"why":"The External Application Load Balancer provides global HTTP(S) load balancing and integrates with Cloud CDN."},{"id":"11","label":"Q11","set":"core","topicId":"designing-and-planning","text":"An internal microservices platform requires service-to-service traffic only inside Google’s network. Which load balancer type is most suitable?","options":["External passthrough NLB","Regional internal Application Load Balancer","Cloud CDN","Internet NEG"],"answer":[1],"why":"An internal Application Load Balancer is designed for private internal HTTP(S) traffic."},{"id":"12","label":"Q12","set":"core","topicId":"designing-and-planning","text":"A platform team wants infrastructure defined in code with repeatable deployments and policy checks. Which approach is best?","options":["Manually click resources in console","Use Terraform with CI/CD","Use spreadsheets to track configs","Only use startup scripts"],"answer":[1],"why":"Infrastructure as code with Terraform enables repeatable, reviewable, and automatable deployments."},{"id":"13","label":"Q13","set":"core","topicId":"designing-and-planning","text":"A company wants to modernize a monolith gradually and expose services through a managed API layer. Which service should be added?","options":["API Gateway or Apigee","Cloud DNS","Cloud Scheduler","Artifact Registry"],"answer":[0],"why":"API Gateway or Apigee can front services, apply policies, and support gradual modernization."},{"id":"14","label":"Q14","set":"core","topicId":"designing-and-planning","text":"A web app has large static assets used globally. Which design change most reduces latency and origin load?","options":["Put assets in Cloud Storage behind Cloud CDN","Move everything to Cloud SQL","Use only regional storage","Enable serial requests"],"answer":[0],"why":"Serving static assets from Cloud Storage with Cloud CDN reduces latency and offloads origin traffic."},{"id":"15","label":"Q15","set":"core","topicId":"designing-and-planning","text":"A company needs a managed Kubernetes platform but wants Google to manage node operations as much as possible. What should it choose?","options":["GKE Autopilot","GKE Standard with self-managed nodes","Compute Engine unmanaged VMs","Cloud Run jobs"],"answer":[0],"why":"GKE Autopilot reduces Kubernetes node-management overhead while keeping Kubernetes compatibility."},{"id":"16","label":"Q16","set":"core","topicId":"designing-and-planning","text":"A team has batch Spark jobs and wants minimal changes during migration. Which service is best?","options":["Dataproc","Cloud SQL","Cloud Run","Firestore"],"answer":[0],"why":"Dataproc is a managed Spark/Hadoop service suited for lifting existing batch data-processing jobs."},{"id":"17","label":"Q17","set":"core","topicId":"designing-and-planning","text":"A global app needs low-latency read access close to users, but writes are relatively infrequent and centralized. Which pattern fits best?","options":["Single-region database with read replicas where supported","One VM in each zone","Only local SSD","Store everything in Secret Manager"],"answer":[0],"why":"Read replicas or distributed read patterns reduce read latency while keeping writes centralized when appropriate."},{"id":"18","label":"Q18","set":"core","topicId":"designing-and-planning","text":"A company wants a hybrid architecture connecting on-prem and Google Cloud privately with predictable bandwidth. Which choice is best?","options":["Cloud VPN over the internet","Dedicated Interconnect","Public IP peering only","Cloud CDN"],"answer":[1],"why":"Dedicated Interconnect provides private connectivity and predictable bandwidth for hybrid environments."},{"id":"19","label":"Q19","set":"core","topicId":"designing-and-planning","text":"A team wants to run ad hoc containers to completion on a schedule, without managing servers. Which service fits best?","options":["Cloud Run jobs with Cloud Scheduler","Compute Engine MIG","GKE StatefulSet","Bigtable"],"answer":[0],"why":"Cloud Run jobs combined with Cloud Scheduler support serverless scheduled container execution."},{"id":"20","label":"Q20","set":"core","topicId":"designing-and-planning","text":"A company has many APIs and needs developer portal, quota plans, monetization, and advanced policy control. Which service is best?","options":["API Gateway","Apigee","Cloud Armor","Cloud Build"],"answer":[1],"why":"Apigee is the full API management platform for advanced lifecycle, policy, and portal needs."},{"id":"21","label":"Q21","set":"core","topicId":"designing-and-planning","text":"An app requires sub-millisecond access to frequently used key-value session data. Which service is best?","options":["Memorystore","Cloud SQL","Cloud Storage","BigQuery"],"answer":[0],"why":"Memorystore offers managed Redis or Memcached for very low-latency caching and session storage."},{"id":"22","label":"Q22","set":"core","topicId":"designing-and-planning","text":"A company wants to categorize data lakes, apply governance, and discover assets across analytics services. Which service helps most?","options":["Dataplex","Deployment Manager","Cloud Shell","Cloud Load Balancing"],"answer":[0],"why":"Dataplex is designed for data discovery, governance, and management across analytics environments."},{"id":"23","label":"Q23","set":"core","topicId":"provisioning-and-infrastructure","text":"A company wants all new projects to inherit baseline IAM, networking constraints, and audit settings. What should it use?","options":["Manual setup per project","Organization policies and automation","Only labels","Individual user scripts"],"answer":[1],"why":"Organization policies plus automation provide scalable baseline governance for new projects."},{"id":"24","label":"Q24","set":"core","topicId":"provisioning-and-infrastructure","text":"A team needs private Google access from VMs without external IPs. Which feature should be enabled?","options":["Private Google Access","Cloud CDN","Cloud Armor","Traffic Director"],"answer":[0],"why":"Private Google Access lets VMs without public IPs reach Google APIs and services."},{"id":"25","label":"Q25","set":"core","topicId":"provisioning-and-infrastructure","text":"A workload in private subnets needs outbound internet access for patching, but should not accept inbound connections. What should you deploy?","options":["Cloud NAT","External IP on each VM","Cloud Armor","Private Service Connect endpoint"],"answer":[0],"why":"Cloud NAT provides outbound internet access without assigning external IPs to instances."},{"id":"26","label":"Q26","set":"core","topicId":"provisioning-and-infrastructure","text":"A company needs a shared VPC model where a networking team owns the VPC and app teams deploy into separate service projects. What should be implemented?","options":["Shared VPC","Peering only","One VPC per VM","Cloud DNS private zone only"],"answer":[0],"why":"Shared VPC enables centralized network administration with decentralized service-project usage."},{"id":"27","label":"Q27","set":"core","topicId":"provisioning-and-infrastructure","text":"A team wants golden VM images for faster standard deployments. Which service or feature should it use?","options":["Instance templates or machine images","BigQuery reservations","Cloud Trace","Datastream"],"answer":[0],"why":"Instance templates and machine images support repeatable, standardized VM deployment patterns."},{"id":"28","label":"Q28","set":"core","topicId":"provisioning-and-infrastructure","text":"A managed instance group must replace unhealthy instances automatically and scale based on CPU. Which setup is correct?","options":["Use an unmanaged instance group","Use MIG with health checks and autoscaling","Use one large VM","Use local SSD only"],"answer":[1],"why":"A managed instance group with health checks and autoscaling is the standard pattern."},{"id":"29","label":"Q29","set":"core","topicId":"provisioning-and-infrastructure","text":"You need a private managed service connection to a producer service without exposing public IPs. Which pattern is best?","options":["Private Service Connect","Cloud VPN","Cloud CDN","Cloud Load Balancer only"],"answer":[0],"why":"Private Service Connect supports private consumption of services over internal IPs."},{"id":"30","label":"Q30","set":"core","topicId":"provisioning-and-infrastructure","text":"A company needs DNS names for internal services inside a VPC. Which solution is best?","options":["Cloud DNS private zone","Public DNS only","Hard-code IPs","Use a spreadsheet"],"answer":[0],"why":"Cloud DNS private zones provide internal DNS resolution within authorized VPCs."},{"id":"31","label":"Q31","set":"core","topicId":"provisioning-and-infrastructure","text":"A platform team wants container images stored privately with vulnerability scanning and IAM controls. Which service is best?","options":["Artifact Registry","Cloud Storage public bucket","Container Registry only on Docker Hub","Filestore"],"answer":[0],"why":"Artifact Registry is the current Google service for storing and managing container images and packages."},{"id":"32","label":"Q32","set":"core","topicId":"provisioning-and-infrastructure","text":"A team must deploy infrastructure across multiple regions with consistent CI/CD. Which design is best?","options":["Manual console deployments","IaC pipeline with environment promotion","Single local script on a laptop","Only snapshots"],"answer":[1],"why":"Using an IaC pipeline with promotion stages ensures consistent multi-region deployments."},{"id":"33","label":"Q33","set":"core","topicId":"provisioning-and-infrastructure","text":"A company wants to bootstrap projects, networks, and service accounts automatically when a new business unit is onboarded. What should it build?","options":["Cloud Functions or Cloud Run automation triggered by workflow","Manual ticket queue","Only organization labels","A wiki page"],"answer":[0],"why":"Automated provisioning reduces error and enforces standards at onboarding."},{"id":"34","label":"Q34","set":"core","topicId":"provisioning-and-infrastructure","text":"A workload needs block storage with very high IOPS attached to a single VM. Which storage type is best?","options":["Persistent Disk or Hyperdisk depending workload needs","Cloud Storage bucket","Filestore share","BigQuery slot"],"answer":[0],"why":"Block storage for a single VM should use managed disk products, selecting the performance tier that matches IOPS needs."},{"id":"35","label":"Q35","set":"core","topicId":"provisioning-and-infrastructure","text":"A company wants blue/green VM deployments with minimal risk. Which approach is best?","options":["Create a second MIG and switch traffic after validation","Delete the old MIG first","Update VMs manually one by one","Use only snapshots"],"answer":[0],"why":"Blue/green with separate managed instance groups enables validation before traffic cutover."},{"id":"36","label":"Q36","set":"core","topicId":"provisioning-and-infrastructure","text":"A GKE team wants cluster access controlled by IAM and short-lived credentials. Which feature is most relevant?","options":["Identity-aware cluster access / Workload Identity and IAM-integrated controls","Static SSH keys","Store kubeconfig in a shared drive","Open cluster to internet"],"answer":[0],"why":"Use IAM-integrated access and short-lived identities rather than shared static credentials."},{"id":"37","label":"Q37","set":"core","topicId":"provisioning-and-infrastructure","text":"A company must use private worker pools for builds accessing internal resources. Which service feature should be used?","options":["Cloud Build private pools","Public runners on random VMs","Cloud CDN signed URLs","Cloud SQL read replica"],"answer":[0],"why":"Cloud Build private pools allow builds to run with controlled network access."},{"id":"38","label":"Q38","set":"core","topicId":"provisioning-and-infrastructure","text":"A team needs workflow orchestration across Google APIs, approvals, and service calls. Which service is best?","options":["Workflows","Cloud CDN","Cloud SQL","Cloud DNS"],"answer":[0],"why":"Workflows orchestrates multi-step processes across services and APIs."},{"id":"39","label":"Q39","set":"core","topicId":"provisioning-and-infrastructure","text":"A company wants asynchronous task execution with rate limiting and deduplication for background HTTP handlers. Which service fits best?","options":["Cloud Tasks","Pub/Sub Lite","Bigtable","Cloud Armor"],"answer":[0],"why":"Cloud Tasks is purpose-built for controlled asynchronous task delivery to HTTP targets."},{"id":"40","label":"Q40","set":"core","topicId":"provisioning-and-infrastructure","text":"A team must provision service accounts with least privilege for each app. Which practice is best?","options":["One shared owner account for all apps","Dedicated service accounts per workload with narrow roles","Use developer personal accounts in prod","Disable IAM review"],"answer":[1],"why":"Dedicated service accounts with least-privilege roles reduce blast radius and improve traceability."},{"id":"41","label":"Q41","set":"core","topicId":"provisioning-and-infrastructure","text":"A company needs secrets like API keys and DB passwords stored centrally and accessed securely by workloads. Which service is best?","options":["Secret Manager","Cloud Storage bucket","BigQuery table","Startup script comments"],"answer":[0],"why":"Secret Manager is the managed service for secrets storage, access control, and rotation integration."},{"id":"42","label":"Q42","set":"core","topicId":"provisioning-and-infrastructure","text":"A team needs a cron-like trigger for a serverless endpoint every hour. Which combination is best?","options":["Cloud Scheduler to Cloud Run/Function","MIG autoscaler","Cloud NAT to Pub/Sub","DNS policy"],"answer":[0],"why":"Cloud Scheduler is the managed cron service for triggering serverless endpoints."},{"id":"43","label":"Q43","set":"core","topicId":"provisioning-and-infrastructure","text":"A company wants a single private entry point to Google APIs from on-prem via private connectivity. Which feature helps?","options":["Private Google Access for on-prem through private networking patterns","Cloud CDN","Public IP allowlist only","Cloud Functions"],"answer":[0],"why":"Private access patterns keep traffic on private paths instead of using public internet when supported."},{"id":"44","label":"Q44","set":"core","topicId":"provisioning-and-infrastructure","text":"A platform must standardize policy checks before Terraform applies. Which addition is most useful?","options":["Policy validation in CI/CD with review gates","More admin accounts","Manual post-deploy review only","Bigger VMs"],"answer":[0],"why":"Shift-left policy checks catch violations before infrastructure is created."},{"id":"45","label":"Q45","set":"core","topicId":"security-and-compliance","text":"A healthcare app stores EHR data in a relational database and must use customer-managed keys with scheduled rotation. Which design is best?","options":["Cloud SQL or Spanner with CMEK via Cloud KMS, depending scale needs","Cloud Storage with public URLs","Bigtable with manual passwords only","Default encryption only"],"answer":[0],"why":"The key requirement is a relational service with CMEK through Cloud KMS for customer-managed rotation."},{"id":"46","label":"Q46","set":"core","topicId":"security-and-compliance","text":"A company wants to enforce MFA and context-aware access to the Google Cloud console. Which product is most relevant?","options":["Identity-Aware Proxy / context-aware access controls","Cloud NAT","Cloud CDN","Bigtable replication"],"answer":[0],"why":"Context-aware access and identity-centric controls are the right fit for console and app access hardening."},{"id":"47","label":"Q47","set":"core","topicId":"security-and-compliance","text":"A team needs WAF protection against OWASP risks and geo/IP filtering for internet-facing services. Which service should be added?","options":["Cloud Armor","Cloud SQL","Cloud DNS","Cloud NAT"],"answer":[0],"why":"Cloud Armor provides WAF and edge security policies for external applications."},{"id":"48","label":"Q48","set":"core","topicId":"security-and-compliance","text":"A company wants to prevent creation of public IPs on VMs organization-wide. What should it use?","options":["Organization Policy constraints","Project labels","Cloud Monitoring only","Billing budgets"],"answer":[0],"why":"Organization Policy can centrally restrict specific resource configurations such as public IP creation."},{"id":"49","label":"Q49","set":"core","topicId":"security-and-compliance","text":"A developer needs temporary access to inspect logs in production. Which practice is best?","options":["Grant a narrow role temporarily using just-in-time access","Grant Owner permanently","Share an admin password","Use a generic service account"],"answer":[0],"why":"Temporary least-privilege access is safer and more auditable than permanent broad roles."},{"id":"50","label":"Q50","set":"core","topicId":"security-and-compliance","text":"A company must classify and inspect sensitive data such as SSNs in storage. Which service is most relevant?","options":["Sensitive Data Protection (Cloud DLP)","Cloud CDN","Cloud Shell","BigQuery BI Engine"],"answer":[0],"why":"Sensitive Data Protection identifies and can mask sensitive information."},{"id":"51","label":"Q51","set":"core","topicId":"security-and-compliance","text":"A team needs workload identities on GKE without storing service account keys in pods. Which feature is best?","options":["Workload Identity Federation for GKE / Workload Identity","Bake keys into container images","Store JSON keys in ConfigMaps","Use node default service account for everything"],"answer":[0],"why":"Workload identity avoids long-lived keys in pods and maps workloads to IAM securely."},{"id":"52","label":"Q52","set":"core","topicId":"security-and-compliance","text":"A company wants all admin actions and data access events centrally logged for forensics. Which services should be configured?","options":["Cloud Audit Logs and centralized logging sinks","Only application stdout","Cloud CDN reports only","No logs to reduce cost"],"answer":[0],"why":"Audit logs plus centralized log retention are essential for security investigations and compliance."},{"id":"53","label":"Q53","set":"core","topicId":"security-and-compliance","text":"A platform team wants to monitor misconfigurations and security findings continuously across projects. Which service helps most?","options":["Security Command Center","Memorystore","Cloud Scheduler","Filestore"],"answer":[0],"why":"Security Command Center centralizes findings and posture management."},{"id":"54","label":"Q54","set":"core","topicId":"security-and-compliance","text":"A company wants VPC-level firewall rules based on service account identity rather than IP ranges alone. Is this a good practice?","options":["Yes, where supported it improves intent-based control","No, only IP-based firewalls should be used","No, open all east-west traffic","Only use DNS rules"],"answer":[0],"why":"Identity-based controls can make network policies more maintainable than broad IP rules."},{"id":"55","label":"Q55","set":"core","topicId":"security-and-compliance","text":"A team must rotate encryption keys regularly and disable a compromised key version quickly. Which service supports this?","options":["Cloud KMS","Cloud DNS","Cloud Deploy","Bigtable"],"answer":[0],"why":"Cloud KMS supports key versions, rotation policies, and controlled disable/destroy workflows."},{"id":"56","label":"Q56","set":"core","topicId":"security-and-compliance","text":"A company needs private access to managed databases without traversing the public internet. Which pattern is best?","options":["Private IP for the managed database and private networking","Public IP with allowlist only","CDN in front of database","Store DB in GCS"],"answer":[0],"why":"Use private IP and private networking where available for managed databases."},{"id":"57","label":"Q57","set":"core","topicId":"security-and-compliance","text":"A finance team wants to ensure logs cannot be tampered with by app operators. What is the best approach?","options":["Centralize logs in a dedicated security project with restricted access","Let app operators delete logs","Store logs only locally on VMs","Disable logging"],"answer":[0],"why":"Separation of duties and centralized log storage reduce tampering risk."},{"id":"58","label":"Q58","set":"core","topicId":"security-and-compliance","text":"A company needs signed identity-based access to internal web apps without a VPN. Which product is best?","options":["Identity-Aware Proxy","Cloud Interconnect","Cloud DNS","Cloud Logging"],"answer":[0],"why":"Identity-Aware Proxy can enforce user identity and context before granting access to internal apps."},{"id":"59","label":"Q59","set":"core","topicId":"security-and-compliance","text":"A team wants to keep secrets out of source control and CI variables. What should they do?","options":["Store secrets in Secret Manager and access them at runtime","Commit encrypted files with the key in README","Use hard-coded values","Email secrets to developers"],"answer":[0],"why":"Secrets should be stored centrally and retrieved securely at runtime."},{"id":"60","label":"Q60","set":"core","topicId":"security-and-compliance","text":"A company wants restricted APIs organization-wide so only approved services can be used. Which control should be applied?","options":["Organization Policy and service usage restrictions","Only VPC firewall rules","Bigger IAM roles","Cloud CDN signed cookies"],"answer":[0],"why":"Org-level policy can restrict which services are allowed or how they may be configured."},{"id":"61","label":"Q61","set":"core","topicId":"security-and-compliance","text":"An app needs mutual TLS between selected services on GKE. Which approach is best?","options":["Use a service mesh or managed mTLS pattern","Expose all services publicly","Put keys in images","Disable cert rotation"],"answer":[0],"why":"A service mesh simplifies mTLS, identity, and certificate rotation between services."},{"id":"62","label":"Q62","set":"core","topicId":"security-and-compliance","text":"A company must prove compliance through recurring evidence of configuration. What should it implement?","options":["Policy-as-code, audit logs, and scheduled posture reports","Manual screenshots once a year","No documentation","Only budget alerts"],"answer":[0],"why":"Automated evidence collection and policy enforcement support repeatable compliance."},{"id":"63","label":"Q63","set":"core","topicId":"security-and-compliance","text":"A team needs to share a BigQuery dataset securely across projects without moving data. Which method is best?","options":["Dataset-level IAM or authorized views","Public export to CSV","Share service account keys broadly","Copy data everywhere"],"answer":[0],"why":"BigQuery supports secure cross-project sharing through IAM and authorized access patterns."},{"id":"64","label":"Q64","set":"core","topicId":"security-and-compliance","text":"A company wants to reduce lateral movement in production. Which action helps most?","options":["Segment environments, use least privilege, and minimize shared identities","Use one flat network for all","Grant owner to all SREs","Disable firewalls"],"answer":[0],"why":"Segmentation and least privilege are core controls against lateral movement."},{"id":"65","label":"Q65","set":"core","topicId":"security-and-compliance","text":"A developer wants to test from a laptop using a service account key file kept for years. What is the better alternative?","options":["Use short-lived credentials and federation where possible","Keep the key forever in Downloads","Share the key with teammates","Put the key in Git"],"answer":[0],"why":"Short-lived or federated credentials are safer than long-lived downloaded keys."},{"id":"66","label":"Q66","set":"core","topicId":"security-and-compliance","text":"A company has data residency requirements. Which design principle matters most?","options":["Choose compliant regions and enforce location-aware architecture","Use any region randomly","Rely only on labels","Ignore backups"],"answer":[0],"why":"Region selection and data placement are foundational to residency compliance."},{"id":"67","label":"Q67","set":"core","topicId":"implementation-and-migration","text":"A company wants to move databases to Google Cloud with minimal downtime and ongoing replication until cutover. Which service is most relevant?","options":["Database Migration Service","Cloud CDN","Dataproc","Secret Manager"],"answer":[0],"why":"Database Migration Service supports managed migrations and replication for supported engines."},{"id":"68","label":"Q68","set":"core","topicId":"implementation-and-migration","text":"A team is containerizing a monolith and wants progressive delivery with approvals across dev, test, and prod. Which service helps?","options":["Cloud Deploy","Cloud DNS","Cloud DLP","Bigtable"],"answer":[0],"why":"Cloud Deploy supports progressive deployment pipelines and approvals."},{"id":"69","label":"Q69","set":"core","topicId":"implementation-and-migration","text":"A company needs to migrate large datasets from on-prem to Cloud Storage over the network on a schedule. Which managed service is relevant?","options":["Storage Transfer Service","Cloud Armor","Looker","Cloud Build"],"answer":[0],"why":"Storage Transfer Service is used for managed transfers into Cloud Storage."},{"id":"70","label":"Q70","set":"core","topicId":"implementation-and-migration","text":"A team has PB-scale on-prem data with limited network bandwidth and needs an offline transfer. Which service is best?","options":["Transfer Appliance","Cloud Run","Memorystore","Workflows"],"answer":[0],"why":"Transfer Appliance is designed for large offline data transfers."},{"id":"71","label":"Q71","set":"core","topicId":"implementation-and-migration","text":"A company wants CI/CD for containers with build, test, scan, and deploy stages. Which toolchain is most aligned with Google Cloud services?","options":["Cloud Build + Artifact Registry + Cloud Deploy","Cloud DNS + Pub/Sub + Filestore","BigQuery + Dataflow","Cloud CDN + Armor"],"answer":[0],"why":"This toolchain covers build, artifact storage, and deployment promotion."},{"id":"72","label":"Q72","set":"core","topicId":"implementation-and-migration","text":"A VM migration project needs to move existing VMware workloads to Google Cloud. Which service is most relevant?","options":["Migrate to Virtual Machines","Cloud Composer","BigQuery Omni","Dataplex"],"answer":[0],"why":"Migrate to Virtual Machines helps assess and migrate VMware-based workloads to Compute Engine."},{"id":"73","label":"Q73","set":"core","topicId":"implementation-and-migration","text":"A team wants to move from VM-based apps to containers gradually, starting with build pipelines first. Which first step is most practical?","options":["Containerize the app and store images in Artifact Registry","Adopt Spanner immediately","Replace all apps with BigQuery","Delete the CI pipeline"],"answer":[0],"why":"A practical first step is to standardize builds and container artifacts."},{"id":"74","label":"Q74","set":"core","topicId":"implementation-and-migration","text":"A company needs CDC from transactional databases into analytics. Which service is useful?","options":["Datastream","Cloud DNS","Cloud Interconnect","Filestore"],"answer":[0],"why":"Datastream provides serverless change data capture into Google Cloud destinations."},{"id":"75","label":"Q75","set":"core","topicId":"implementation-and-migration","text":"A data team wants managed orchestration for ETL DAGs using Apache Airflow. Which service should it choose?","options":["Cloud Composer","Cloud Scheduler only","Cloud Deploy","Cloud Shell"],"answer":[0],"why":"Cloud Composer is the managed Apache Airflow service."},{"id":"76","label":"Q76","set":"core","topicId":"implementation-and-migration","text":"A company is replatforming a web app and wants zero-downtime migration between versions. Which pattern is best?","options":["Canary or blue/green deployments behind a load balancer","Big-bang cutover without rollback","Manual SSH updates","No health checks"],"answer":[0],"why":"Canary and blue/green reduce risk and support fast rollback."},{"id":"77","label":"Q77","set":"core","topicId":"implementation-and-migration","text":"A team wants schema-managed streaming ETL with autoscaling and exactly-once style processing semantics where appropriate. Which service is best?","options":["Dataflow","Cloud SQL","GKE Autopilot","Memorystore"],"answer":[0],"why":"Dataflow is the managed stream/batch processing service for Beam pipelines."},{"id":"78","label":"Q78","set":"core","topicId":"implementation-and-migration","text":"A company wants to move an API backend to Cloud Run with minimal client changes. Which addition helps preserve the external contract?","options":["Use API Gateway or load balancer in front of Cloud Run","Force clients to change all URLs immediately","Expose random ports directly","Use Cloud Storage website hosting"],"answer":[0],"why":"An API layer or load balancer can preserve routes, security, and client contract during migration."},{"id":"79","label":"Q79","set":"core","topicId":"implementation-and-migration","text":"A team wants feature flags and gradual release by user percentage. Which approach is best?","options":["Add application-level feature flags and canary delivery","Deploy everything at once","Use DNS changes only","Use a spreadsheet"],"answer":[0],"why":"Feature flags plus gradual rollout reduce release risk."},{"id":"80","label":"Q80","set":"core","topicId":"implementation-and-migration","text":"A company must migrate from self-managed Redis to a managed service with minimal application change. Which target is best?","options":["Memorystore for Redis","Bigtable","Cloud SQL","Cloud Storage"],"answer":[0],"why":"Memorystore for Redis is the managed target with the least change for Redis-style use cases."},{"id":"81","label":"Q81","set":"core","topicId":"implementation-and-migration","text":"A team wants to verify infrastructure drift after manual changes. Which practice is best?","options":["Run IaC plan/validation regularly and reconcile drift","Ignore drift","Delete prod and recreate blindly","Use only local scripts"],"answer":[0],"why":"Regular drift detection maintains consistency between desired and actual infrastructure."},{"id":"82","label":"Q82","set":"core","topicId":"implementation-and-migration","text":"A company has a legacy NFS dependency and wants to modernize later, not now. Which migration landing zone is best?","options":["Compute Engine with Filestore","Cloud Run only","BigQuery","Pub/Sub Lite"],"answer":[0],"why":"Compute Engine plus Filestore supports legacy dependencies while enabling future modernization."},{"id":"83","label":"Q83","set":"core","topicId":"implementation-and-migration","text":"A company wants to package Python libraries for internal use across teams. Which service is appropriate?","options":["Artifact Registry package repositories","Cloud Storage public bucket","Secret Manager","Filestore"],"answer":[0],"why":"Artifact Registry supports package formats beyond containers."},{"id":"84","label":"Q84","set":"core","topicId":"implementation-and-migration","text":"A team wants centralized deployment approvals and release history for GKE and Cloud Run. Which service fits?","options":["Cloud Deploy","Cloud Build triggers only","Cloud Logging","Cloud DNS"],"answer":[0],"why":"Cloud Deploy provides managed release pipelines and approvals."},{"id":"85","label":"Q85","set":"core","topicId":"implementation-and-migration","text":"A company is modernizing a batch scheduler into event-driven automation. Which pair is often a good replacement?","options":["Pub/Sub plus Cloud Run or Functions","Cloud DNS plus Armor","Filestore plus Spanner","Bigtable plus CDN"],"answer":[0],"why":"Event-driven processing on Pub/Sub with serverless handlers is a common modernization path."},{"id":"86","label":"Q86","set":"core","topicId":"implementation-and-migration","text":"A team wants to standardize app configuration separately from container images across environments. Which practice is best?","options":["Externalize config via environment variables or config services and keep images immutable","Build a separate image for each secret","Hard-code prod values","Edit containers manually"],"answer":[0],"why":"Immutable images plus externalized configuration support reliable deployments."},{"id":"87","label":"Q87","set":"core","topicId":"implementation-and-migration","text":"A company wants to rehearse disaster recovery cutovers before migration completes. Which practice is best?","options":["Run game days and validated runbooks regularly","Trust the docs without testing","Wait for a real disaster","Skip monitoring"],"answer":[0],"why":"Practiced DR exercises expose gaps before they become outages."},{"id":"88","label":"Q88","set":"core","topicId":"implementation-and-migration","text":"A platform team wants all releases traceable from commit to deployment. Which design is best?","options":["Use CI/CD with artifact digests, release metadata, and audit trails","Deploy from laptops","Overwrite images with latest only","Skip source control tags"],"answer":[0],"why":"Traceability is essential for change management and rollback confidence."},{"id":"89","label":"Q89","set":"core","topicId":"optimization-and-cost","text":"A workload runs only on weekdays during business hours and uses non-production VMs. What is the easiest cost optimization?","options":["Schedule start/stop for the VMs","Increase machine size","Add more regions","Use premium storage everywhere"],"answer":[0],"why":"Stopping non-prod VMs outside needed hours often yields immediate savings."},{"id":"90","label":"Q90","set":"core","topicId":"optimization-and-cost","text":"A stable baseline compute fleet runs continuously. Which purchasing model can lower cost compared with on-demand pricing?","options":["Committed use discounts where appropriate","Cloud CDN","More snapshots","Premium images only"],"answer":[0],"why":"Committed use discounts reduce cost for predictable steady-state usage."},{"id":"91","label":"Q91","set":"core","topicId":"optimization-and-cost","text":"A batch analytics job can tolerate interruption. Which compute option is often cheapest?","options":["Spot VMs where suitable","Highest-end dedicated hosts always","Multi-year overprovisioning","Memorystore"],"answer":[0],"why":"Spot VMs are discounted for interruptible-tolerant workloads."},{"id":"92","label":"Q92","set":"core","topicId":"optimization-and-cost","text":"A BigQuery team is scanning much more data than needed. Which query design change helps most?","options":["Partition and cluster tables appropriately","Add more views only","Store data in PDFs","Use SELECT * everywhere"],"answer":[0],"why":"Partitioning and clustering can significantly reduce scanned data and cost."},{"id":"93","label":"Q93","set":"core","topicId":"optimization-and-cost","text":"A company stores infrequently accessed backups but must retain them for years at low cost. Which storage class should be considered?","options":["Archive or Coldline as appropriate","Local SSD","Filestore High Scale SSD","Bigtable SSD"],"answer":[0],"why":"Archive-class object storage is designed for low-cost long-term retention."},{"id":"94","label":"Q94","set":"core","topicId":"optimization-and-cost","text":"A team wants to identify idle disks, underutilized VMs, and oversized resources. Which service or capability helps?","options":["Recommender / cost optimization recommendations","Cloud DNS","Artifact Registry","Pub/Sub schemas"],"answer":[0],"why":"Recommendations help surface idle and underutilized resources."},{"id":"95","label":"Q95","set":"core","topicId":"optimization-and-cost","text":"A media app serves global static content but origin egress cost is high. What should it enable?","options":["Cloud CDN","Bigtable replication","Cloud DLP","Cloud Composer"],"answer":[0],"why":"Cloud CDN reduces origin egress and improves user latency."},{"id":"96","label":"Q96","set":"core","topicId":"optimization-and-cost","text":"A company wants separate billing visibility by department without changing the org structure. What should it use?","options":["Labels and billing export to BigQuery","Only one shared spreadsheet","Cloud NAT","Secret Manager"],"answer":[0],"why":"Labels plus billing export support departmental chargeback and analysis."},{"id":"97","label":"Q97","set":"core","topicId":"optimization-and-cost","text":"A data processing job has bursty demand. Which platform choice best avoids paying for idle capacity?","options":["Serverless options such as Cloud Run or Dataflow autoscaling where suitable","Large fixed VM clusters","Dedicated hardware all year","Filestore"],"answer":[0],"why":"Serverless and autoscaling services reduce idle infrastructure spend."},{"id":"98","label":"Q98","set":"core","topicId":"optimization-and-cost","text":"A company has many small Cloud SQL instances with low utilization. Which change might reduce cost?","options":["Consolidate where security and performance requirements allow","Increase all instance sizes","Add more replicas for no reason","Export to PDFs"],"answer":[0],"why":"Consolidation can reduce fixed instance overhead when tenancy requirements allow."},{"id":"99","label":"Q99","set":"core","topicId":"optimization-and-cost","text":"A team needs frequent analytics on operational data without overloading the OLTP database. What is best?","options":["Replicate/stream data to an analytics store such as BigQuery","Run all BI on primary OLTP","Use bigger passwords","Disable indexes"],"answer":[0],"why":"Separating OLTP and analytics improves performance and can lower operational risk."},{"id":"100","label":"Q100","set":"core","topicId":"optimization-and-cost","text":"A company wants alerts when spend exceeds thresholds. What should it configure?","options":["Billing budgets and alerts","Cloud Armor policies","Cloud NAT log sampling","Cloud DNS peering"],"answer":[0],"why":"Budgets and alerts provide proactive spend visibility."},{"id":"101","label":"Q101","set":"core","topicId":"optimization-and-cost","text":"A team is repeatedly rebuilding unchanged dependencies in CI, slowing builds and raising cost. Which optimization helps?","options":["Use build caching and layered images","Add more manual steps","Disable CI","Use larger PDFs"],"answer":[0],"why":"Caching cuts build time and resource consumption."},{"id":"102","label":"Q102","set":"core","topicId":"optimization-and-cost","text":"A global service needs low latency but not all data must be stored in the expensive highest-performance tier. What design can reduce cost?","options":["Tier storage and data paths by access pattern","Keep everything on premium SSD always","Duplicate all data in every region","Store logs in RAM only"],"answer":[0],"why":"Right-tiering based on access pattern balances performance and spend."},{"id":"103","label":"Q103","set":"core","topicId":"optimization-and-cost","text":"A company wants to optimize BigQuery slot usage for a predictable analytics workload. Which option may help?","options":["Capacity planning with reservations where justified","Move everything to Cloud SQL","Use public buckets only","Disable BI tools"],"answer":[0],"why":"Reservations can be economical for sustained predictable workloads."},{"id":"104","label":"Q104","set":"core","topicId":"optimization-and-cost","text":"A team is logging huge volumes of low-value debug data in production. Which action is best?","options":["Reduce log verbosity and tune exclusions/retention carefully","Log even more","Disable all logs","Store logs in source code"],"answer":[0],"why":"Reducing low-value logs cuts storage cost and signal noise while preserving critical telemetry."},{"id":"105","label":"Q105","set":"core","topicId":"optimization-and-cost","text":"A company is paying for many public egress transfers between services in different regions. What architecture review is needed?","options":["Review data locality and co-locate chatty services where appropriate","Add more random regions","Increase cache TTL for databases","Use bigger VMs only"],"answer":[0],"why":"Data locality can significantly reduce network cost and latency."},{"id":"106","label":"Q106","set":"core","topicId":"optimization-and-cost","text":"A company uses many images in Artifact Registry but wants to lower storage costs safely. What should it do?","options":["Apply retention and cleanup policies","Delete everything nightly","Push images to public repos","Disable tags"],"answer":[0],"why":"Retention policies remove stale artifacts while keeping needed versions."},{"id":"107","label":"Q107","set":"core","topicId":"optimization-and-cost","text":"A team wants to compare monthly spend trends by service and label. Which setup is best?","options":["Export billing data to BigQuery and build dashboards","Read invoices manually only","Use Secret Manager","Use Cloud Shell history"],"answer":[0],"why":"Billing export enables detailed custom analysis."},{"id":"108","label":"Q108","set":"core","topicId":"optimization-and-cost","text":"A workload has CPU spikes driven by queue depth rather than CPU metrics. Which autoscaling trigger may be more cost-efficient?","options":["Scale on the relevant business metric or queue length","Scale only on memory always","Disable autoscaling","Use one giant VM"],"answer":[0],"why":"Aligning autoscaling to real demand avoids both over- and under-provisioning."},{"id":"109","label":"Q109","set":"core","topicId":"optimization-and-cost","text":"A company wants to minimize storage costs for temporary ETL files that are recreated daily. Which action helps?","options":["Use lifecycle rules to expire them automatically","Store forever in multi-region","Replicate to all regions","Back them up for 7 years"],"answer":[0],"why":"Automatic expiration prevents paying for disposable data."},{"id":"110","label":"Q110","set":"core","topicId":"optimization-and-cost","text":"A team wants to know the total cost impact of a new architecture before launch. Which practice is best?","options":["Estimate with pricing tools and model expected traffic/storage","Guess based on intuition","Deploy first and hope","Skip performance tests"],"answer":[0],"why":"Cost modeling before launch helps avoid expensive surprises."},{"id":"111","label":"Q111","set":"core","topicId":"reliability-and-operations","text":"A mission-critical service must survive a zonal failure in one region. What is the minimum recommended design?","options":["Deploy across multiple zones with regional load balancing where appropriate","Use one VM in one zone","Back up weekly only","Store docs in Drive"],"answer":[0],"why":"Multi-zone architecture is the baseline for zonal failure resilience."},{"id":"112","label":"Q112","set":"core","topicId":"reliability-and-operations","text":"An application needs disaster recovery against regional outage with a low RTO and moderate cost sensitivity. Which pattern is often suitable?","options":["Warm standby in a second region","Tape backup only","Single-zone design","No runbooks"],"answer":[0],"why":"Warm standby balances recovery time and cost better than active-active for many workloads."},{"id":"113","label":"Q113","set":"core","topicId":"reliability-and-operations","text":"A team wants SLO-based monitoring and alerting focused on user experience. What should it track first?","options":["Golden signals and SLIs tied to user journeys","Only CPU temperature","Number of emails sent","Color of dashboards"],"answer":[0],"why":"User-centric SLIs/SLOs make alerting meaningful and actionable."},{"id":"114","label":"Q114","set":"core","topicId":"reliability-and-operations","text":"A platform team wants distributed tracing across microservices to diagnose latency. Which service is relevant?","options":["Cloud Trace","Cloud DNS","Cloud Storage","Artifact Registry"],"answer":[0],"why":"Cloud Trace helps analyze request latency across distributed systems."},{"id":"115","label":"Q115","set":"core","topicId":"reliability-and-operations","text":"A team needs structured log queries and centralized retention. Which service is best?","options":["Cloud Logging","Cloud CDN","Cloud SQL","Cloud DLP"],"answer":[0],"why":"Cloud Logging centralizes and queries operational logs."},{"id":"116","label":"Q116","set":"core","topicId":"reliability-and-operations","text":"A company wants metrics dashboards and alert policies for infrastructure and applications. Which service is appropriate?","options":["Cloud Monitoring","Cloud Spanner","Cloud Source Repositories","Secret Manager"],"answer":[0],"why":"Cloud Monitoring is the managed telemetry and alerting platform."},{"id":"117","label":"Q117","set":"core","topicId":"reliability-and-operations","text":"A stateful database requires backups with point-in-time recovery. What should be configured?","options":["Automated backups and PITR where supported","Snapshots once a year only","No backups because it is managed","Public exports only"],"answer":[0],"why":"Backups and PITR are core operational safeguards for databases."},{"id":"118","label":"Q118","set":"core","topicId":"reliability-and-operations","text":"A team wants runbooks triggered automatically when alerts fire. Which design is best?","options":["Monitoring alerts integrated with automation workflows","Only manual paging with no docs","Email everyone every minute","Disable alerts"],"answer":[0],"why":"Alert-driven automation shortens response time and standardizes remediation."},{"id":"119","label":"Q119","set":"core","topicId":"reliability-and-operations","text":"An API must stay available during instance restarts. Which pattern helps most?","options":["Multiple instances behind a load balancer with health checks","One large instance only","Manual failover scripts only","Cron restarts"],"answer":[0],"why":"Redundancy behind health-checked load balancers avoids single-instance outages."},{"id":"120","label":"Q120","set":"core","topicId":"reliability-and-operations","text":"A team wants to verify backups are usable. What should it do?","options":["Run periodic restore tests","Trust backup success logs only","Disable encryption","Delete old copies immediately"],"answer":[0],"why":"Only restore testing proves that backups are recoverable."},{"id":"121","label":"Q121","set":"core","topicId":"reliability-and-operations","text":"A service uses Pub/Sub subscriptions and processing occasionally falls behind. What is a key operational metric to watch?","options":["Subscription backlog and processing latency","Bucket object count only","DNS TTL only","SSH key age only"],"answer":[0],"why":"Backlog and latency reveal whether consumers are keeping up."},{"id":"122","label":"Q122","set":"core","topicId":"reliability-and-operations","text":"A company wants error budgets to drive release pace. What does that mean?","options":["Release faster when service reliability is within budget; slow changes when budget is burned","Ignore outages","Only track cost","Deploy only yearly"],"answer":[0],"why":"Error budgets connect reliability targets to change velocity."},{"id":"123","label":"Q123","set":"core","topicId":"reliability-and-operations","text":"A data pipeline must continue processing if one worker fails. Which design is best?","options":["Use managed distributed processing with checkpoint/retry semantics","Use one VM","Write results only at the end with no retries","Disable health checks"],"answer":[0],"why":"Distributed managed processing with retries/checkpoints improves resiliency."},{"id":"124","label":"Q124","set":"core","topicId":"reliability-and-operations","text":"A team wants to reduce alert fatigue. Which change helps most?","options":["Tune actionable alerts around symptoms and severity","Alert on every metric by default","Page for every log line","Disable dashboards"],"answer":[0],"why":"Actionable, symptom-based alerting reduces noise."},{"id":"125","label":"Q125","set":"core","topicId":"reliability-and-operations","text":"A globally used read-heavy service needs resilience and low latency. Which pattern helps?","options":["Use global load balancing and replicate/cache reads appropriately","Serve all users from one zone","Use one static VM","No cache"],"answer":[0],"why":"Global frontends plus distributed read paths improve resilience and latency."},{"id":"126","label":"Q126","set":"core","topicId":"reliability-and-operations","text":"A company wants postmortems that improve learning rather than blame. Which practice is best?","options":["Blameless postmortems with actions and owners","Punish the on-call engineer","Avoid documentation","Only fix symptoms"],"answer":[0],"why":"Blameless postmortems encourage honest learning and systemic fixes."},{"id":"127","label":"Q127","set":"core","topicId":"reliability-and-operations","text":"A platform team needs deployment rollback within minutes. Which design choice matters most?","options":["Immutable versioned artifacts and automated rollback paths","Manual prod edits","Single mutable server","No release history"],"answer":[0],"why":"Versioned immutable releases make rollback predictable and fast."},{"id":"128","label":"Q128","set":"core","topicId":"reliability-and-operations","text":"A service has hard external dependencies. How can reliability improve?","options":["Add timeouts, retries with backoff, and circuit breaker patterns","Retry forever instantly","Block all user requests","Ignore failures"],"answer":[0],"why":"Controlled retries and circuit breaking prevent cascading failures."},{"id":"129","label":"Q129","set":"core","topicId":"reliability-and-operations","text":"A batch system must rerun safely after partial failure. Which job property is most valuable?","options":["Idempotency","Maximum logging only","Randomized output paths","Manual approval every record"],"answer":[0],"why":"Idempotent jobs can rerun without duplicating side effects."},{"id":"130","label":"Q130","set":"core","topicId":"reliability-and-operations","text":"A team wants operational visibility into application performance by endpoint. Which telemetry is most useful?","options":["Request latency, error rate, throughput, and saturation by service/endpoint","Only disk serial number","DNS registrar info","Container image size only"],"answer":[0],"why":"These core service metrics expose endpoint-level health and performance."},{"id":"131","label":"Q131","set":"core","topicId":"reliability-and-operations","text":"A company needs a DR objective discussion with the business. Which pair of metrics should be agreed first?","options":["RTO and RPO","CPU and RAM","Price and color","Zone and label"],"answer":[0],"why":"Recovery time objective and recovery point objective are the essential DR targets."},{"id":"132","label":"Q132","set":"core","topicId":"reliability-and-operations","text":"A platform serves traffic globally and wants controlled failover between regions. Which design is best?","options":["Global load balancing with healthy regional backends and tested failover","Manual DNS edits during outage","One region only","No health checks"],"answer":[0],"why":"Health-aware global routing supports reliable regional failover."},{"id":"133","label":"Q133","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat Media streams content globally and traffic spikes during live events. Which frontend design is best?","options":["External Application Load Balancer with autoscaling backends and Cloud CDN","One zonal VM","Internal load balancer only","Static HTML on a laptop"],"answer":[0],"why":"Live-event traffic benefits from global frontends, autoscaling, and edge caching."},{"id":"134","label":"Q134","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat stores media assets and needs low-latency global delivery. Which service should front origin storage?","options":["Cloud CDN","Cloud SQL","Cloud KMS","Cloud Composer"],"answer":[0],"why":"Cloud CDN is the right choice for caching globally distributed media assets."},{"id":"135","label":"Q135","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants to process uploaded videos asynchronously after ingestion. Which event-driven pattern is best?","options":["Upload to Cloud Storage -> Pub/Sub/Eventarc -> Cloud Run/Dataflow processing","Process synchronously on the web tier","Write directly to BigQuery first","Use DNS failover"],"answer":[0],"why":"Event-driven processing decouples upload from compute-intensive transcoding."},{"id":"136","label":"Q136","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat needs metadata search and analytics on viewer behavior at scale. Which storage split is best?","options":["Operational metadata in a suitable app database; analytics in BigQuery","Put everything in Cloud DNS","Keep analytics in Memorystore","Use only local files"],"answer":[0],"why":"Operational and analytical workloads should be separated by access pattern."},{"id":"137","label":"Q137","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants CI/CD for frequent feature releases to a containerized app. Which managed toolchain is best?","options":["Cloud Build + Artifact Registry + Cloud Deploy","Manual SSH","Cron jobs only","Filestore snapshots"],"answer":[0],"why":"This toolchain supports repeatable container build and release workflows."},{"id":"138","label":"Q138","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat has global users and wants user uploads to land closest to users while keeping backend simple. Which choice helps?","options":["Use signed uploads to Cloud Storage with global frontend routing","Tunnel uploads through one VM","Force FTP to one region","Email attachments"],"answer":[0],"why":"Signed direct uploads reduce app-tier bottlenecks and scale better globally."},{"id":"139","label":"Q139","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants origin shielding from abusive traffic during a popular release. Which additional control is best?","options":["Cloud Armor","Cloud Trace","Cloud DNS","Cloud Scheduler"],"answer":[0],"why":"Cloud Armor protects internet-facing endpoints against abusive or malicious traffic."},{"id":"140","label":"Q140","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat needs to transcode many videos in parallel using existing containerized workers. Which compute platform is a strong fit if jobs run to completion?","options":["Cloud Run jobs or GKE batch pattern depending control needs","Cloud SQL read replicas","Cloud DLP","Cloud CDN only"],"answer":[0],"why":"Run-to-completion container jobs fit serverless jobs or batch-oriented container platforms."},{"id":"141","label":"Q141","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat’s recommendation engine needs near-real-time event ingestion from user clicks. Which service starts the stream best?","options":["Pub/Sub","Cloud Interconnect","Filestore","Artifact Registry"],"answer":[0],"why":"Pub/Sub is the standard event-ingestion backbone for near-real-time click streams."},{"id":"142","label":"Q142","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants personalized content recommendations trained on large historical interaction data. Which analytics stack is most suitable?","options":["BigQuery plus ML/Vertex-related analytics patterns as needed","Cloud DNS plus NAT","Filestore plus NFS","Compute Engine only"],"answer":[0],"why":"Large-scale historical analysis is best handled in the analytics stack rather than the serving tier."},{"id":"143","label":"Q143","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat needs signed access to premium content objects. Which capability helps most?","options":["Signed URLs or signed cookies","Public buckets","Store links in source code","Disable IAM"],"answer":[0],"why":"Signed access provides time-bound controlled access to premium assets."},{"id":"144","label":"Q144","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat expects some live event processing to be interruptible for cost savings. Which compute option should be considered for fault-tolerant batch parts?","options":["Spot VMs where interruption is acceptable","Single expensive VM only","Dedicated hosts for everything","Archive storage"],"answer":[0],"why":"Spot capacity can lower cost for fault-tolerant workers."},{"id":"145","label":"Q145","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat’s ad-serving microservice must respond in milliseconds and cache hot lookup data. Which addition helps most?","options":["Memorystore","Cloud Composer","BigQuery export","Transfer Appliance"],"answer":[0],"why":"Hot ad or profile lookup data belongs in a fast cache."},{"id":"146","label":"Q146","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants central observability across all microservices. Which pair is essential?","options":["Cloud Logging and Cloud Monitoring","Cloud DNS and Cloud Storage","Cloud KMS and Secrets only","Gmail and Calendar"],"answer":[0],"why":"Logs plus metrics are the operational baseline for distributed services."},{"id":"147","label":"Q147","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat must support regional outages with limited budget. Which DR pattern is often reasonable?","options":["Warm standby in a second region","No DR","Only zone redundancy","Manual rebuild from memory"],"answer":[0],"why":"Warm standby gives meaningful resilience without full active-active cost."},{"id":"148","label":"Q148","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants to separate production, staging, and analytics access boundaries. What should it do?","options":["Use separate projects/environments with IAM separation","Put everything in one project as Owner","Share one service account everywhere","Disable audit logs"],"answer":[0],"why":"Environment isolation simplifies access control and risk management."},{"id":"149","label":"Q149","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat uploads grow quickly and older assets become less frequently accessed. What storage optimization should it apply?","options":["Lifecycle rules to tier older objects","Keep all objects on hottest class forever","Delete everything after 7 days","Move to local SSD"],"answer":[0],"why":"Lifecycle transitions align storage cost with access frequency."},{"id":"150","label":"Q150","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat needs to ensure user profile PII is protected in logs and datasets. Which service helps inspect and mask sensitive data?","options":["Sensitive Data Protection","Cloud CDN","Cloud NAT","Artifact Registry"],"answer":[0],"why":"Sensitive Data Protection is the right service for PII discovery and masking."},{"id":"151","label":"Q151","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants rapid frontend deployment for static landing pages. Which hosting pattern is best?","options":["Cloud Storage with load balancer/CDN or Firebase-style managed hosting pattern","Cloud SQL","Bigtable","Filestore NFS"],"answer":[0],"why":"Static hosting with edge delivery is the simplest and fastest approach."},{"id":"152","label":"Q152","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat plans to expose partner APIs for content metadata. They need quotas, auth policies, and developer onboarding. Which product fits best?","options":["Apigee","Cloud DNS","Cloud Spanner","Cloud Shell"],"answer":[0],"why":"Apigee is designed for managed partner/API product exposure."},{"id":"153","label":"Q153","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat must analyze the cost of traffic surges by release and geography. Which setup provides the best flexibility?","options":["Billing export to BigQuery with labels/tags and custom dashboards","Monthly invoice screenshots only","Cloud Trace only","Secret Manager"],"answer":[0],"why":"Billing export supports custom cost analysis by dimensions such as label and service."},{"id":"154","label":"Q154","set":"core","topicId":"case-study-altostrat-media","text":"Altostrat wants to minimize build times for many microservices. Which CI optimization is most useful?","options":["Layer caching and shared base images with Cloud Build","Manual builds on laptops","Disable tests","Store binaries in email"],"answer":[0],"why":"Caching and shared base layers reduce CI time and spend."},{"id":"155","label":"Q155","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal Retail expects seasonal spikes and wants a highly scalable web storefront with low ops. Which frontend is best?","options":["Cloud Run behind External Application Load Balancer","Single VM","Cloud SQL only","Local desktop server"],"answer":[0],"why":"A serverless web tier behind global load balancing handles bursty retail traffic with low ops overhead."},{"id":"156","label":"Q156","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal’s product catalog changes less frequently than shopping cart state. Which storage pattern is sensible?","options":["Cache catalog aggressively and store cart state in a low-latency operational store","Keep cart data in Cloud Storage objects","Put everything in DNS","Store orders in RAM only"],"answer":[0],"why":"Catalog and cart have different access patterns and should be treated differently."},{"id":"157","label":"Q157","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal needs search over product attributes and analytics on sales trends. What is a good split?","options":["Operational catalog in app DB/search layer; analytics in BigQuery","BigQuery for live cart transactions only","DNS for attributes","Filestore for BI"],"answer":[0],"why":"Separate operational commerce paths from analytical reporting."},{"id":"158","label":"Q158","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants fraud-detection events from checkout to trigger downstream review workflows. Which eventing design is best?","options":["Publish checkout events to Pub/Sub and process asynchronously","Block checkout while every downstream system responds","Write events into comments","Use DNS updates"],"answer":[0],"why":"Asynchronous eventing prevents tight coupling on the checkout path."},{"id":"159","label":"Q159","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal needs PCI-minded tokenization and secure secret handling. Which two services are most relevant?","options":["Tokenization pattern plus Secret Manager/KMS controls","Cloud CDN and DNS","Dataproc and Transfer Appliance","Only labels"],"answer":[0],"why":"Payments require secure tokenization and strong secret/key management."},{"id":"160","label":"Q160","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants to personalize offers using customer behavior in near real time. Which ingestion service starts the pipeline best?","options":["Pub/Sub","Filestore","Cloud DNS","Cloud Shell"],"answer":[0],"why":"Pub/Sub is the common starting point for real-time event ingestion."},{"id":"161","label":"Q161","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal must keep checkout available even if recommendations fail. Which design principle should they apply?","options":["Decouple recommendations from checkout and fail gracefully","Make checkout depend on the recommender","Block on every API call","Turn off health checks"],"answer":[0],"why":"Critical purchase flows should degrade gracefully when noncritical services fail."},{"id":"162","label":"Q162","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal uses GKE and wants pod identities without managing keys. Which feature should it use?","options":["Workload Identity","Long-lived JSON keys in secrets","Node default identity for all apps","Shared admin token"],"answer":[0],"why":"Workload Identity is the secure choice for GKE workloads."},{"id":"163","label":"Q163","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants a managed relational database for orders with automatic failover. Which service is usually the straightforward choice when scale is moderate?","options":["Cloud SQL with HA configuration","Bigtable","Cloud Storage","Secret Manager"],"answer":[0],"why":"For moderate-scale relational orders, Cloud SQL HA is often the pragmatic managed choice."},{"id":"164","label":"Q164","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal’s order pipeline needs exactly-once downstream invoicing behavior. Which application design is most important?","options":["Idempotent consumers and deduplication logic","Infinite retries with no keys","One huge server","Manual spreadsheets"],"answer":[0],"why":"Idempotency is the key safeguard for event-driven financial processes."},{"id":"165","label":"Q165","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants mobile app images to load quickly worldwide. Which addition is most useful?","options":["Cloud CDN","Cloud KMS","Cloud Monitoring","Dataproc"],"answer":[0],"why":"Edge caching improves mobile asset performance globally."},{"id":"166","label":"Q166","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants to isolate supplier integrations from the core retail platform. Which design principle should they choose?","options":["Loosely coupled APIs/events with clear contracts","Direct shared database access for suppliers","FTP into production DB","Open all firewall ports"],"answer":[0],"why":"Suppliers should integrate through controlled contracts, not shared internal state."},{"id":"167","label":"Q167","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal needs dashboards for revenue by region and product category. Which service fits best?","options":["BigQuery with BI dashboards","Cloud DNS","Memorystore","Filestore"],"answer":[0],"why":"BigQuery is the natural analytics backend for revenue dashboards."},{"id":"168","label":"Q168","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants progressive delivery for storefront changes during holiday season. Which deployment strategy is safest?","options":["Canary release with automated rollback","Big-bang cutover","Manual SSH edits","No metrics"],"answer":[0],"why":"Canary rollout reduces risk during peak revenue periods."},{"id":"169","label":"Q169","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal needs WAF and bot protection on the storefront. Which service should be configured?","options":["Cloud Armor","Cloud Run jobs","Secret Manager","Cloud SQL Insights"],"answer":[0],"why":"Cloud Armor protects the public storefront from common web threats."},{"id":"170","label":"Q170","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal must support omnichannel inventory updates from stores and online systems. Which architecture pattern fits best?","options":["Event-driven inventory updates with Pub/Sub and downstream consumers","Nightly manual CSV only","One spreadsheet as source of truth","DNS-based inventory"],"answer":[0],"why":"Event-driven updates improve freshness and decoupling across channels."},{"id":"171","label":"Q171","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants logs partitioned so security analysts have access but store managers do not. What should it do?","options":["Centralize logs and enforce IAM by role and project","Email logs to everyone","Disable logging","Store logs in buckets with public access"],"answer":[0],"why":"Centralized logging with role-based IAM supports separation of duties."},{"id":"172","label":"Q172","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal needs low-latency session storage for web users. Which service is best?","options":["Memorystore for Redis","BigQuery","Cloud Storage","Cloud DLP"],"answer":[0],"why":"Sessions are a classic low-latency cache/store use case."},{"id":"173","label":"Q173","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal expects high cart abandonment and wants event-driven remarketing exports. Which service can schedule recurring export workflows easily?","options":["Workflows or Scheduler-triggered jobs","Cloud DNS peering","Filestore snapshots","Cloud NAT"],"answer":[0],"why":"Managed workflow/scheduling is suitable for recurring export orchestration."},{"id":"174","label":"Q174","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal must comply with data retention rules for order logs. Which capability helps automate retention?","options":["Log buckets/retention settings and storage lifecycle policies","Manual deletion reminders","Disable logs","Put all logs in local files"],"answer":[0],"why":"Retention controls automate compliance for logs and stored records."},{"id":"175","label":"Q175","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal wants to optimize cloud spend after peak season. Which first action is best?","options":["Review recommender findings, rightsize, and shut down non-prod excess","Migrate everything again","Increase every quota","Add more regions"],"answer":[0],"why":"Post-peak right-sizing often yields immediate savings."},{"id":"176","label":"Q176","set":"core","topicId":"case-study-cymbal-retail","text":"Cymbal’s analytics team wants curated domains for sales, inventory, and marketing. Which service helps govern the data lake/warehouse estate?","options":["Dataplex","Cloud Armor","Compute Engine","Transfer Appliance"],"answer":[0],"why":"Dataplex helps organize and govern analytics domains."},{"id":"177","label":"Q177","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR Healthcare needs HIPAA-aligned architecture and customer-managed key rotation for regulated data. Which key service is essential?","options":["Cloud KMS","Cloud CDN","Cloud Trace","Cloud Scheduler"],"answer":[0],"why":"Cloud KMS provides customer-managed encryption keys and rotation controls."},{"id":"178","label":"Q178","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR stores patient records transactionally with strong relational requirements and moderate scale. Which database is often the most straightforward fit?","options":["Cloud SQL with private IP and CMEK where supported/required by design choice","Bigtable","Cloud Storage only","Memorystore"],"answer":[0],"why":"For many transactional healthcare apps, a managed relational database is the pragmatic choice."},{"id":"179","label":"Q179","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR must separate PHI processing from public portal traffic. Which network design is best?","options":["Use separate tiers/projects/VPC segmentation with private service access for backend systems","Put everything in one flat public subnet","Expose DB publicly","Disable firewalls"],"answer":[0],"why":"Segmenting PHI systems from public tiers reduces risk and supports least privilege."},{"id":"180","label":"Q180","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants clinicians to access an internal app without a VPN, but with strong identity checks. Which product is best?","options":["Identity-Aware Proxy","Cloud CDN","Cloud NAT","BigQuery"],"answer":[0],"why":"Identity-Aware Proxy adds identity-based access to internal apps."},{"id":"181","label":"Q181","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR needs to identify PHI appearing in logs or exports. Which service helps most?","options":["Sensitive Data Protection","Cloud Monitoring","Cloud DNS","Artifact Registry"],"answer":[0],"why":"Sensitive Data Protection is used for detecting and masking sensitive content."},{"id":"182","label":"Q182","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR must ensure every admin action is auditable. Which feature is essential?","options":["Cloud Audit Logs","Cloud Shell themes","Local log files only","No logging to reduce storage"],"answer":[0],"why":"Audit Logs are essential for compliance and forensic traceability."},{"id":"183","label":"Q183","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants least-privilege access for separate teams handling app support, security, and analytics. What should it do?","options":["Use IAM role separation and dedicated service accounts","Grant Owner to all teams","Share a root password","Use one generic service account"],"answer":[0],"why":"Role separation and workload-specific identities are foundational controls."},{"id":"184","label":"Q184","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR needs to move HL7/FHIR-related messages between systems asynchronously. Which service is a good backbone for decoupled messaging?","options":["Pub/Sub","Cloud Storage static site","Filestore","Cloud DNS"],"answer":[0],"why":"Pub/Sub is the standard asynchronous messaging backbone."},{"id":"185","label":"Q185","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR has a patient portal front end that must withstand internet attacks. Which service should front it?","options":["Cloud Armor","Cloud Composer","Bigtable","Transfer Appliance"],"answer":[0],"why":"Cloud Armor provides WAF and edge protection for internet-facing apps."},{"id":"186","label":"Q186","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR requires private connectivity from on-prem hospital systems with predictable performance. Which option is best if usage is substantial?","options":["Dedicated Interconnect","Public internet only","Cloud CDN","Public bucket sync"],"answer":[0],"why":"Dedicated Interconnect is preferred for substantial private hybrid connectivity needs."},{"id":"187","label":"Q187","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants encrypted secrets for application config with rotation support. Which service is best?","options":["Secret Manager","Cloud Storage","Memorystore","Cloud Build"],"answer":[0],"why":"Secret Manager provides centralized controlled secret storage."},{"id":"188","label":"Q188","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR must minimize blast radius if a service account is compromised. Which practice is best?","options":["Use separate narrow service accounts per workload and short-lived credentials","Use one broad admin account","Store keys in Git","Reuse the same key forever"],"answer":[0],"why":"Workload-specific least-privilege identities limit blast radius."},{"id":"189","label":"Q189","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR needs high availability against zonal failure for the patient portal. Which architecture is minimum baseline?","options":["Multi-zone deployment behind a load balancer","One VM in one zone","Nightly backup only","Desktop app"],"answer":[0],"why":"Multi-zone redundancy is the baseline protection against zonal failure."},{"id":"190","label":"Q190","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants a full DR plan for regional outage with tested failover. Which business metrics should be documented first?","options":["RTO and RPO","CPU and disk","Project and folder","MTU and TTL"],"answer":[0],"why":"RTO and RPO guide the DR architecture and testing decisions."},{"id":"191","label":"Q191","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR must store imaging files cost-effectively for long retention. Which storage design is appropriate?","options":["Cloud Storage with lifecycle tiering and retention controls","Local SSD only","Memorystore","BigQuery BI Engine"],"answer":[0],"why":"Cloud Storage supports durable object storage and lifecycle-based cost control."},{"id":"192","label":"Q192","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants to prevent accidental public exposure of buckets and public IPs org-wide. What should be used?","options":["Organization Policy constraints","Bigger VMs","DNS naming rules only","One giant firewall rule"],"answer":[0],"why":"Org policies centrally restrict risky configurations."},{"id":"193","label":"Q193","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants centralized security posture findings across projects. Which service helps most?","options":["Security Command Center","Cloud Trace","Cloud Scheduler","Datastream"],"answer":[0],"why":"Security Command Center centralizes findings and posture visibility."},{"id":"194","label":"Q194","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR needs asynchronous document processing for scanned forms uploaded by clinics. Which serverless pattern is good?","options":["Cloud Storage trigger -> Pub/Sub/Eventarc -> Cloud Run/Document pipeline","Synchronous processing in user request","Manual email processing","Store in DNS"],"answer":[0],"why":"Decoupled serverless processing is scalable and resilient for document workflows."},{"id":"195","label":"Q195","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants separate analytics on de-identified data without impacting the transactional system. Which target is most suitable?","options":["BigQuery with de-identification pipeline","Run all analytics on primary Cloud SQL","Use local CSV only","Store data in Secret Manager"],"answer":[0],"why":"De-identified analytics belong in the warehouse, not on the transactional primary."},{"id":"196","label":"Q196","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR must preserve data residency for certain patient datasets. What is the key architecture action?","options":["Choose approved regions and enforce location-aware storage/processing","Use any default region","Only label the data","Ignore backups"],"answer":[0],"why":"Region selection and placement controls are essential for residency compliance."},{"id":"197","label":"Q197","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR wants zero trust access to internal admin apps. Which pattern is aligned?","options":["Identity-centric access with IAP and context-aware controls","Flat VPN for everyone with shared password","Public app plus basic auth","Disable MFA"],"answer":[0],"why":"Identity-centric access is aligned with zero-trust principles."},{"id":"198","label":"Q198","set":"core","topicId":"case-study-ehr-healthcare","text":"EHR needs evidence for auditors that backups and restores work. What should operations do?","options":["Run periodic restore drills and document results","Trust backup success messages","Never test to avoid risk","Store audits in email"],"answer":[0],"why":"Auditors care that recovery is proven, not assumed."},{"id":"199","label":"Q199","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives Automotive collects telemetry from connected vehicles worldwide. Which ingestion service is the best starting point for decoupled streaming?","options":["Pub/Sub","Cloud DNS","Cloud CDN","Filestore"],"answer":[0],"why":"Pub/Sub is the standard decoupled ingestion service for event streams."},{"id":"200","label":"Q200","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives needs very high write throughput for time-series telemetry keyed by vehicle and timestamp. Which database fits best?","options":["Bigtable","Cloud SQL","Cloud Storage website","Secret Manager"],"answer":[0],"why":"Bigtable is well-suited for large-scale time-series telemetry workloads."},{"id":"201","label":"Q201","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants fleet analytics and historical trend analysis over large datasets using SQL. Which service is best?","options":["BigQuery","Memorystore","Filestore","Cloud NAT"],"answer":[0],"why":"BigQuery is the right service for large-scale analytical SQL over historical telemetry."},{"id":"202","label":"Q202","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives must process telemetry in near real time to detect anomalies. Which service is most suitable?","options":["Dataflow","Cloud DNS","Cloud Armor","Cloud SQL"],"answer":[0],"why":"Dataflow supports scalable real-time stream processing."},{"id":"203","label":"Q203","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants over-the-air update packages distributed globally to vehicles and service centers. Which edge service helps delivery performance?","options":["Cloud CDN","Cloud KMS","Cloud Build private pools","Cloud Composer"],"answer":[0],"why":"Cloud CDN improves global distribution of static update packages."},{"id":"204","label":"Q204","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives has multiple supplier APIs and wants a governed interface for partners. Which service is best?","options":["Apigee","Cloud Shell","Transfer Appliance","Bigtable backup"],"answer":[0],"why":"Apigee provides the partner-facing API management capabilities."},{"id":"205","label":"Q205","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants to isolate critical manufacturing systems from analytics workloads. Which organizational pattern is best?","options":["Separate projects/environments with IAM and network segmentation","Single flat project for everything","One shared owner account","No logs"],"answer":[0],"why":"Segmentation reduces blast radius and clarifies governance."},{"id":"206","label":"Q206","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants event-driven processing per region but centralized policy control. Which design is sensible?","options":["Regional ingestion with standardized pipelines and central governance","One VM in headquarters","Manual CSV uploads","Only spreadsheets"],"answer":[0],"why":"Regional pipelines reduce latency while central governance keeps standards consistent."},{"id":"207","label":"Q207","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives needs anomaly alerts sent to ops tools when stream thresholds are crossed. Which architecture fits best?","options":["Dataflow/stream processing publishes alert events to downstream systems","Daily batch reports only","DNS failover","Local cron on laptop"],"answer":[0],"why":"Real-time stream processing is needed for anomaly-driven alerting."},{"id":"208","label":"Q208","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants to reduce cost for noncritical simulation jobs. Which compute option should it consider?","options":["Spot VMs","Dedicated hosts only","Single giant VM","Filestore SSD"],"answer":[0],"why":"Simulation jobs that tolerate interruption are good candidates for Spot VMs."},{"id":"209","label":"Q209","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives must protect vehicle identity and owner data in analytics exports. Which service helps detect and mask sensitive data?","options":["Sensitive Data Protection","Cloud Monitoring","Artifact Registry","Cloud Scheduler"],"answer":[0],"why":"Sensitive Data Protection is used for discovery and de-identification of sensitive data."},{"id":"210","label":"Q210","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants reliable ingestion even during short downstream outages. Which buffering pattern helps most?","options":["Pub/Sub decoupling with durable subscriptions","Direct synchronous writes everywhere","FTP only","DNS-based buffering"],"answer":[0],"why":"Durable subscriptions absorb short consumer outages without dropping events."},{"id":"211","label":"Q211","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants SLOs around telemetry freshness. Which SLI is most relevant?","options":["End-to-end event processing latency","Bucket count","Number of IAM roles","VM boot time"],"answer":[0],"why":"Freshness is measured by how quickly an event becomes available/processed end to end."},{"id":"212","label":"Q212","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives must share a subset of analytics securely with a supplier. Which option is best?","options":["BigQuery dataset sharing or authorized views","Public CSV in a bucket","Share admin credentials","Open firewall to database"],"answer":[0],"why":"Authorized data sharing avoids unnecessary copying and overexposure."},{"id":"213","label":"Q213","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants immutable deployment artifacts for vehicle-update backend services. Which service should store the images?","options":["Artifact Registry","Filestore","Cloud Storage public bucket","Cloud Trace"],"answer":[0],"why":"Artifact Registry stores container artifacts with IAM and scanning support."},{"id":"214","label":"Q214","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives needs secure secret delivery for microservices in GKE. Which service is best?","options":["Secret Manager","Cloud DNS","Bigtable","Cloud NAT"],"answer":[0],"why":"Secret Manager centralizes secret storage and controlled access."},{"id":"215","label":"Q215","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants highly available APIs across zones in one region. Which compute pattern is baseline?","options":["Multiple instances or pods across zones behind a load balancer","One zonal VM","Manual failover once a week","Desktop app"],"answer":[0],"why":"Multi-zone redundancy behind a load balancer is the baseline."},{"id":"216","label":"Q216","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants cost reports by vehicle program and geography. Which setup helps most?","options":["Billing export to BigQuery with labels/tags","Only PDF invoices","Cloud Trace spans","Cloud DLP scans"],"answer":[0],"why":"Detailed billing export supports custom cost attribution."},{"id":"217","label":"Q217","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives needs private access from manufacturing plants to Google APIs without public IP exposure on workloads. Which feature is helpful?","options":["Private Google Access combined with private networking","Cloud CDN","Public buckets","Cloud Armor only"],"answer":[0],"why":"Private Google Access supports API access from VMs without external IPs."},{"id":"218","label":"Q218","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants to coordinate a multi-step remediation flow after a critical alert. Which service is suited for this orchestration?","options":["Workflows","Cloud DNS","Memorystore","Filestore"],"answer":[0],"why":"Workflows coordinates multi-step automated responses."},{"id":"219","label":"Q219","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives wants to archive older telemetry cheaply but still retain it for compliance. Which storage action helps?","options":["Lifecycle policies to colder Cloud Storage classes","Keep all data in hot storage forever","Delete after 30 days","Store only on local SSD"],"answer":[0],"why":"Lifecycle policies reduce retention cost for older telemetry."},{"id":"220","label":"Q220","set":"core","topicId":"case-study-knightmotives-automotive","text":"KnightMotives needs a managed dashboard of security findings across projects. Which service provides this?","options":["Security Command Center","Cloud Build","Cloud SQL","Cloud Run jobs"],"answer":[0],"why":"Security Command Center centralizes security findings and posture insights."},{"id":"s1","label":"SQ1","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A retail company wants to build a product recommendation system. The team has ML expertise and needs custom model training with managed infrastructure, experiment tracking, and model serving. Which service is the best fit?","options":["Cloud Composer","Vertex AI","BigQuery only","Dataflow"],"answer":[1],"why":"Vertex AI provides end-to-end ML lifecycle management including training, experiment tracking, and model serving. Cloud Composer orchestrates workflows but isn't an ML platform. BigQuery ML handles simpler models but lacks full lifecycle support for custom training."},{"id":"s2","label":"SQ2","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A logistics company wants to extract text from scanned shipping labels without training a custom model. Which approach requires the least effort?","options":["Train a custom model on Vertex AI","Use the Cloud Vision API","Build a Dataflow pipeline with TensorFlow","Use BigQuery ML"],"answer":[1],"why":"Cloud Vision API is a pre-trained API for OCR and image analysis—no model training needed. Vertex AI custom training and Dataflow+TF are overengineered for this. BigQuery ML doesn't handle image processing."},{"id":"s3","label":"SQ3","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A financial services company wants to use generative AI to summarize internal research documents but must keep data within its Google Cloud tenant. Which approach is most appropriate?","options":["Use the public Gemini consumer app","Deploy Vertex AI with Gemini models and configure VPC Service Controls","Fine-tune an open-source model on a personal laptop","Store documents in Pub/Sub and query them"],"answer":[1],"why":"Vertex AI with Gemini provides enterprise-grade GenAI with data residency controls, IAM, and VPC Service Controls for data protection. The consumer app offers no enterprise data guarantees."},{"id":"s4","label":"SQ4","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A team needs to serve a custom TensorFlow model with autoscaling, A/B traffic splitting, and monitoring. Which is the best option?","options":["Deploy on a single Compute Engine VM","Use Cloud Run with the model bundled in a container","Use Vertex AI Endpoints","Serve from Cloud Storage signed URLs"],"answer":[2],"why":"Vertex AI Endpoints provide managed model serving with autoscaling, traffic splitting for A/B testing, and integrated monitoring. Cloud Run could work for simpler cases but lacks native ML serving features like traffic splitting between model versions."},{"id":"s5","label":"SQ5","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"An e-commerce team wants to add product image search without ML expertise. Users upload a photo and the app finds similar catalog items. Which approach is most practical? (Select TWO)","options":["Vertex AI Vision with image similarity search","Write a custom CNN from scratch on Compute Engine","Use the Vision API for label detection plus a vector search","Store images in BigQuery and use SQL LIKE queries"],"answer":[0,2],"why":"Vertex AI Vision and the Vision API combined with vector search are both viable managed approaches for image similarity. Building a custom CNN from scratch requires deep ML expertise. BigQuery SQL doesn't support pixel-level image search."},{"id":"s6","label":"SQ6","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A healthcare company uses BigQuery for patient analytics and wants to add predictive models without moving data outside BigQuery or managing ML infrastructure. Which is the best choice?","options":["Export to Cloud Storage and train on Vertex AI","BigQuery ML","Dataproc Spark ML","Cloud Functions with scikit-learn"],"answer":[1],"why":"BigQuery ML trains models directly inside BigQuery using SQL—no data export or separate infrastructure needed. While Vertex AI offers more advanced ML capabilities, it requires data movement, which the requirements explicitly exclude."},{"id":"s7","label":"SQ7","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A DevOps team wants to classify incoming support tickets by urgency and route them automatically. They have 50,000 labeled historical tickets. What is the most efficient approach?","options":["AutoML text classification on Vertex AI","Write regex rules manually","Use Cloud Natural Language API entity extraction","Build a custom BERT model on bare Compute Engine"],"answer":[0],"why":"With labeled data and a classification task, AutoML text classification on Vertex AI trains a custom model with minimal ML expertise. The NL API provides general analysis but not custom classification. Manual regex won't scale or generalize."},{"id":"s8","label":"SQ8","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A media company wants to moderate user-uploaded video content at scale. Which combination provides the most comprehensive automated approach?","options":["Video Intelligence API for content moderation","Manual human review only","Cloud DLP on video metadata","Store videos in Bigtable and scan labels"],"answer":[0],"why":"The Video Intelligence API can detect explicit content, violence, and other moderation signals in video at scale. Cloud DLP inspects text/structured data, not video content directly."},{"id":"s9","label":"SQ9","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A company grounding a GenAI chatbot on internal knowledge bases wants to reduce hallucination and restrict answers to verified company data. Which Vertex AI feature is most relevant?","options":["Vertex AI Search and Conversation with grounding","Train a model from scratch on all company data","Use Pub/Sub to stream documents to the model","Fine-tune Gemini with RLHF only"],"answer":[0],"why":"Vertex AI Search and Conversation with grounding connects GenAI models to enterprise data stores, restricting responses to verified content and reducing hallucination. Full retraining is expensive and unnecessary. RLHF alone doesn't ground responses in specific documents."},{"id":"s10","label":"SQ10","set":"advanced","topicId":"ai-ml-and-generative-ai","text":"A team deploying ML models needs to ensure only cryptographically verified container images are used in production. Which service enforces this?","options":["Cloud Armor","Binary Authorization","Cloud NAT","Artifact Registry tags only"],"answer":[1],"why":"Binary Authorization enforces deploy-time policy by requiring images to be signed by trusted authorities before they can run on GKE or Cloud Run. Artifact Registry stores images but doesn't enforce runtime deployment policies."},{"id":"s11","label":"SQ11","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A financial institution must prevent data exfiltration from BigQuery datasets even if an IAM identity is compromised. Which security layer adds this protection?","options":["VPC Service Controls","Cloud CDN signed cookies","Cloud NAT","Organization policy for public IP only"],"answer":[0],"why":"VPC Service Controls create a security perimeter around GCP services like BigQuery, preventing data from leaving the perimeter even if IAM credentials are compromised. This is the primary tool for data exfiltration prevention."},{"id":"s12","label":"SQ12","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A company configures VPC Service Controls and needs to allow a trusted third-party SaaS to access a BigQuery dataset inside the perimeter. What should they configure?","options":["Remove the perimeter temporarily","An ingress rule in the service perimeter","A Cloud VPN tunnel to the SaaS","Public dataset access"],"answer":[1],"why":"Ingress rules allow specific approved external identities to access resources inside a VPC Service Controls perimeter without removing protection."},{"id":"s13","label":"SQ13","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A company wants to enforce that Cloud Storage objects can only be accessed from within specific VPC networks and by approved identities. Which combination provides the strongest control?","options":["IAM plus VPC Service Controls","ACLs only","Signed URLs only","Object versioning"],"answer":[0],"why":"IAM controls identity, and VPC Service Controls restrict the network context from which API calls can be made. Together they provide defense in depth. ACLs and signed URLs don't restrict network-level access context."},{"id":"s14","label":"SQ14","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A security team discovers that a service account key was leaked externally. Which immediate actions should they take? (Select TWO)","options":["Disable the leaked key version in IAM","Review Cloud Audit Logs for unauthorized usage","Increase the VM machine type","Delete all unrelated projects"],"answer":[0,1],"why":"The immediate response is to disable the compromised credential and audit what actions were taken with it. Reviewing audit logs identifies the scope of the breach."},{"id":"s15","label":"SQ15","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A company needs custom organization policy constraints beyond the built-in options. Which feature allows this?","options":["Custom organization policy constraints","IAM deny policies only","Cloud Armor custom rules","VPC firewall rules"],"answer":[0],"why":"Custom organization policy constraints let administrators define validation rules for resource configurations beyond what built-in constraints offer."},{"id":"s16","label":"SQ16","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A company operating in the EU must ensure that support personnel outside the EU cannot access customer data. Which feature enforces this access control?","options":["Access Transparency and Access Approval","Cloud CDN geo-restriction","Billing alerts","Cloud DNS DNSSEC"],"answer":[0],"why":"Access Transparency logs Google personnel access, and Access Approval requires explicit customer approval before Google support can access data—key for sovereignty and compliance requirements."},{"id":"s17","label":"SQ17","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A GKE-based application processes financial transactions and the security team requires that no unsigned container images run in the production cluster. Which is the best approach?","options":["Use naming conventions for images","Configure Binary Authorization with attestation policies","Scan images monthly with a cron job","Allow only public Docker Hub images"],"answer":[1],"why":"Binary Authorization enforces that only attested (cryptographically signed) images can be deployed, providing a hard policy gate—not just convention or periodic scanning."},{"id":"s18","label":"SQ18","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"An enterprise wants to federate identities from their on-premises Active Directory to Google Cloud without syncing passwords. Which solution is correct?","options":["Google Cloud Directory Sync (GCDS) plus SAML SSO federation","Create new Google accounts for every user manually","Share a single service account with all users","Use Cloud Functions to replicate passwords"],"answer":[0],"why":"GCDS syncs user and group information from AD to Cloud Identity, and SAML SSO federation handles authentication—passwords never leave the on-premises IdP."},{"id":"s19","label":"SQ19","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A company wants IAM policies that explicitly deny certain permissions, overriding any allow policies. Which feature supports this?","options":["IAM deny policies","Organization policy constraints","VPC firewall deny rules","Cloud Armor deny rules"],"answer":[0],"why":"IAM deny policies provide explicit deny rules that take precedence over allow policies at any level of the resource hierarchy. Organization policies restrict resource configurations, not IAM permissions."},{"id":"s20","label":"SQ20","set":"advanced","topicId":"vpc-service-controls-and-advanced-security","text":"A company needs to manage and rotate TLS certificates automatically for its Global External Application Load Balancer. Which service should it use?","options":["Manually upload certificates weekly","Certificate Manager with Google-managed certificates","Store certificates in Secret Manager only","Cloud DNS TXT records"],"answer":[1],"why":"Certificate Manager provides automated provisioning and renewal of TLS certificates for load balancers, eliminating manual rotation and reducing the risk of expired certificates."},{"id":"s21","label":"SQ21","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A company runs Kubernetes clusters both on-premises and in Google Cloud. They want a single management plane for fleet-wide policy, configuration, and observability. Which product fits best?","options":["GKE Enterprise (Anthos)","Cloud Run on each environment separately","Compute Engine managed instance groups","Cloud Composer"],"answer":[0],"why":"GKE Enterprise (formerly Anthos) provides a unified management plane across on-prem, Google Cloud, and other cloud Kubernetes clusters for policy, config sync, and observability."},{"id":"s22","label":"SQ22","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A platform team using GKE Enterprise wants to enforce consistent Kubernetes resource configurations (namespaces, RBAC, network policies) across 15 clusters. Which component should they use?","options":["Config Sync","Cloud Scheduler","Terraform only","Cloud Build"],"answer":[0],"why":"Config Sync is the GKE Enterprise component that syncs Kubernetes configs from a Git repo to multiple clusters, ensuring consistent policy enforcement."},{"id":"s23","label":"SQ23","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A company runs workloads on AWS EKS and Google GKE and wants unified service mesh observability and mTLS across both. Which approach is best?","options":["Anthos Service Mesh / GKE Enterprise Service Mesh","Set up VPN tunnels only","Use Cloud CDN for both","Run nginx proxies on each cluster"],"answer":[0],"why":"Anthos Service Mesh (now part of GKE Enterprise) provides cross-cluster and cross-cloud mTLS, traffic management, and unified observability for multi-cloud Kubernetes."},{"id":"s24","label":"SQ24","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A company wants to run standard Kubernetes workloads on-premises using Google-managed infrastructure software. Which deployment model fits?","options":["GKE on-prem (part of GKE Enterprise)","Cloud Run on bare metal","Cloud Functions on-prem","Compute Engine on-prem"],"answer":[0],"why":"GKE on-prem (part of GKE Enterprise) brings Google-managed Kubernetes to on-premises data centers with consistent APIs and tooling."},{"id":"s25","label":"SQ25","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A team using GKE needs to enforce that pods in the \"payments\" namespace cannot communicate with pods in the \"analytics\" namespace. Which Kubernetes resource is most appropriate?","options":["Network Policy","IAM role binding","Cloud Armor policy","VPC firewall rule"],"answer":[0],"why":"Kubernetes Network Policies control pod-to-pod traffic at the namespace and label level. IAM and Cloud Armor operate at different layers. VPC firewalls can't distinguish Kubernetes namespaces."},{"id":"s26","label":"SQ26","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A company needs GKE clusters that automatically adjust the number and type of nodes based on workload requirements without managing node pools. Which mode should they choose?","options":["GKE Standard with manual node pools","GKE Autopilot","Compute Engine unmanaged instances","Cloud Run"],"answer":[1],"why":"GKE Autopilot manages nodes automatically, including right-sizing and scaling, based on workload resource requests—no manual node pool management needed."},{"id":"s27","label":"SQ27","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A GKE cluster runs both GPU-intensive ML training and standard API workloads. How should you isolate these workload types?","options":["Use separate node pools with taints/tolerations and appropriate machine types","Run everything on one default node pool","Use separate GKE clusters per container","Only use CPU for all workloads"],"answer":[0],"why":"Separate node pools allow GPU-accelerated machines for ML and standard machines for APIs. Taints and tolerations ensure workloads are scheduled to the correct pool."},{"id":"s28","label":"SQ28","set":"advanced","topicId":"anthos-gke-enterprise-multi-cloud","text":"A company running GKE wants to expose an internal microservice only to other services within the cluster using gRPC. Which resource type is most appropriate?","options":["Ingress with external IP","ClusterIP Service","NodePort Service","External Application Load Balancer"],"answer":[1],"why":"ClusterIP is the default Kubernetes Service type for internal-only cluster communication. It doesn't expose the service externally. Ingress and NodePort expose services outside the cluster."},{"id":"s29","label":"SQ29","set":"advanced","topicId":"advanced-networking","text":"A company needs private connectivity from on-premises to Google Cloud at 50 Gbps with an SLA. Which option meets the requirement?","options":["Cloud VPN","Partner Interconnect","Dedicated Interconnect with multiple 10G or 100G links","Public internet with HTTPS"],"answer":[2],"why":"Dedicated Interconnect supports 10 Gbps or 100 Gbps physical connections with SLA-backed uptime. Cloud VPN is limited to ~3 Gbps per tunnel. Partner Interconnect maxes out at lower capacity than 50 Gbps reliably."},{"id":"s30","label":"SQ30","set":"advanced","topicId":"advanced-networking","text":"An organization uses Shared VPC and wants a service project's GKE cluster to use IP ranges from the host project's VPC. What must be configured?","options":["Secondary IP ranges on the host project subnet for pods and services","Create a new VPC in the service project","Use VPC peering instead","Enable Cloud CDN on the host project"],"answer":[0],"why":"GKE in Shared VPC requires secondary IP ranges on the host project subnet allocated for pod and service CIDRs. The service project doesn't need its own VPC—that defeats the purpose of Shared VPC."},{"id":"s31","label":"SQ31","set":"advanced","topicId":"advanced-networking","text":"A company needs VPN connectivity to Google Cloud but requires an SLA and doesn't want to manage physical cross-connects. Which option is best?","options":["Classic VPN","HA VPN","Dedicated Interconnect","Partner Interconnect"],"answer":[3],"why":"Partner Interconnect provides SLA-backed private connectivity through a service provider without the customer needing to manage physical infrastructure. HA VPN runs over the public internet (no private SLA). Dedicated Interconnect requires physical cross-connects."},{"id":"s32","label":"SQ32","set":"advanced","topicId":"advanced-networking","text":"A team running services across two VPCs in different projects needs bidirectional communication without overlapping CIDR ranges. Which is the simplest approach?","options":["VPC peering","Deploy a Cloud VPN between them","Use Shared VPC and merge into one VPC","Use Cloud CDN"],"answer":[0],"why":"VPC peering provides direct, low-latency, high-bandwidth communication between two VPCs without deploying additional infrastructure. CIDRs must not overlap. Shared VPC restructuring may not be needed for just two VPCs."},{"id":"s33","label":"SQ33","set":"advanced","topicId":"advanced-networking","text":"A company's on-premises DNS resolver must resolve names for resources in Google Cloud private DNS zones. Which configuration is needed?","options":["DNS forwarding from Cloud DNS to on-prem","Inbound DNS forwarding policy in Cloud DNS","Create public DNS records for everything","Hard-code IPs in /etc/hosts on-prem"],"answer":[1],"why":"An inbound DNS forwarding policy creates a forwarding target in the VPC that on-premises DNS resolvers can query to resolve Cloud DNS private zone records over private connectivity."},{"id":"s34","label":"SQ34","set":"advanced","topicId":"advanced-networking","text":"A network team wants to analyze VPC traffic patterns for security investigations. Which feature should be enabled?","options":["VPC Flow Logs","Cloud CDN logs","Pub/Sub audit logs","Binary Authorization logs"],"answer":[0],"why":"VPC Flow Logs record network flows to and from VM interfaces, providing data for security analysis, forensics, and network monitoring."},{"id":"s35","label":"SQ35","set":"advanced","topicId":"advanced-networking","text":"A company needs to choose between VPC peering and Shared VPC. Which factor favors Shared VPC over peering?","options":["Need for centralized network administration with decentralized workload deployment","Connecting two VPCs in different organizations","Temporary test connectivity","Exactly two VPCs that never change"],"answer":[0],"why":"Shared VPC is designed for centralized network governance where a networking team manages the VPC while application teams deploy into service projects. VPC peering is simpler for connecting independent VPCs but doesn't centralize administration."},{"id":"s36","label":"SQ36","set":"advanced","topicId":"advanced-networking","text":"A company needs their internal services to be accessible across multiple regions with automatic failover and a single anycast IP. Which load balancing option supports this for internal clients?","options":["Cross-region internal Application Load Balancer","Regional external passthrough NLB","Cloud CDN","Cloud DNS round-robin"],"answer":[0],"why":"The cross-region internal Application Load Balancer provides a single internal IP that routes traffic across regions with automatic failover—ideal for multi-region internal services."},{"id":"s37","label":"SQ37","set":"advanced","topicId":"advanced-database-comparison","text":"An application needs a globally distributed relational database with strong consistency, automatic sharding, and the ability to handle hundreds of thousands of reads/writes per second. Which is the best fit?","options":["Cloud SQL with read replicas","Cloud Spanner","AlloyDB","Firestore"],"answer":[1],"why":"Cloud Spanner is the only managed relational database on Google Cloud that provides global distribution, automatic horizontal sharding, and strongly consistent transactions at scale. Cloud SQL doesn't auto-shard. AlloyDB is regional."},{"id":"s38","label":"SQ38","set":"advanced","topicId":"advanced-database-comparison","text":"A team migrating a PostgreSQL-heavy workload needs the highest possible performance for mixed transactional and analytical queries, but only within a single region. Which database should they evaluate?","options":["Cloud SQL for PostgreSQL","Cloud Spanner","AlloyDB for PostgreSQL","Bigtable"],"answer":[2],"why":"AlloyDB is a PostgreSQL-compatible, high-performance database designed for demanding transactional and analytical workloads within a single region. It outperforms standard Cloud SQL for PostgreSQL and doesn't require schema changes. Spanner requires migration from PostgreSQL dialect."},{"id":"s39","label":"SQ39","set":"advanced","topicId":"advanced-database-comparison","text":"A mobile gaming application needs a serverless document database that syncs data directly to mobile clients with offline support. Which database is the best fit?","options":["Firestore (Native mode)","Cloud SQL","Cloud Spanner","Bigtable"],"answer":[0],"why":"Firestore in Native mode provides real-time sync, offline support for mobile/web clients, and serverless scaling. No other listed database offers native mobile client SDKs with offline sync."},{"id":"s40","label":"SQ40","set":"advanced","topicId":"advanced-database-comparison","text":"A company needs to store 500 TB of time-series IoT data with single-millisecond read latency for key lookups but does NOT need SQL joins or complex transactions. Which database is best?","options":["Cloud Spanner","Cloud SQL for PostgreSQL","BigQuery","Bigtable"],"answer":[3],"why":"Bigtable excels at massive-scale time-series data with single-digit millisecond latency for key-based lookups. It doesn't support SQL joins or ACID transactions, which aren't needed here. BigQuery is for analytics, not low-latency lookups."},{"id":"s41","label":"SQ41","set":"advanced","topicId":"advanced-database-comparison","text":"When should an architect recommend Cloud Spanner over AlloyDB? (Select TWO)","options":["The workload requires multi-region strong consistency","The workload needs automatic horizontal sharding beyond single-node capacity","The team wants the cheapest option for a small single-region database","The application uses PostgreSQL stored procedures heavily"],"answer":[0,1],"why":"Spanner's advantages over AlloyDB are global distribution with strong consistency and unlimited horizontal sharding. AlloyDB is typically more cost-efficient for single-region PostgreSQL workloads and supports full PostgreSQL compatibility including stored procedures."},{"id":"s42","label":"SQ42","set":"advanced","topicId":"advanced-database-comparison","text":"An analytics team runs complex SQL joins over 50 TB of historical data, primarily for dashboards refreshed hourly. Which storage/query engine is most appropriate?","options":["Bigtable","Memorystore","BigQuery","Firestore"],"answer":[2],"why":"BigQuery is the serverless analytics warehouse built for complex SQL over massive datasets. Bigtable doesn't support SQL joins. Memorystore is for caching. Firestore is a document database."},{"id":"s43","label":"SQ43","set":"advanced","topicId":"advanced-database-comparison","text":"A company needs to migrate an Oracle OLTP database to Google Cloud. The app uses many Oracle-specific PL/SQL features. Which approach minimizes application code changes?","options":["Rewrite for Cloud Spanner","Use Bare Metal Solution for Oracle","Migrate to Firestore","Use Bigtable"],"answer":[1],"why":"Bare Metal Solution allows running Oracle workloads on dedicated hardware in Google Cloud data centers with minimal application changes. All other options require significant code rewriting."},{"id":"s44","label":"SQ44","set":"advanced","topicId":"advanced-database-comparison","text":"A team needs to decide between Firestore Native mode and Datastore mode. Which statement is correct?","options":["Datastore mode supports real-time listeners and mobile SDK sync","Native mode supports real-time listeners and mobile/web SDKs; Datastore mode is better for server-only workloads","They are identical in all features","Datastore mode is newer than Native mode"],"answer":[1],"why":"Native mode adds real-time sync, offline support, and mobile/web client libraries. Datastore mode is optimized for server-side applications and backward compatibility with the legacy Datastore API."},{"id":"s45","label":"SQ45","set":"advanced","topicId":"advanced-cost-optimization","text":"A company runs a steady-state GKE cluster 24/7 for 3 years. Which purchasing options should they evaluate to reduce compute cost? (Select TWO)","options":["Committed use discounts (CUDs) for compute","Spot VMs for all production pods","Sustained use discounts (automatic)","Move everything to Cloud Functions"],"answer":[0,2],"why":"CUDs provide the deepest discounts for predictable multi-year usage. Sustained use discounts apply automatically for consistent monthly usage. Spot VMs are inappropriate for always-on production workloads due to preemption risk."},{"id":"s46","label":"SQ46","set":"advanced","topicId":"advanced-cost-optimization","text":"A data team runs BigQuery with highly predictable daily query volume of 500 TB scanned/day. Which pricing model should they evaluate?","options":["On-demand pricing (per TB scanned)","BigQuery Editions with capacity commitments (slots)","Move all queries to Cloud SQL","Use Dataproc instead"],"answer":[1],"why":"At 500 TB/day of predictable usage, capacity-based pricing with slot commitments is typically much more cost-effective than on-demand per-TB pricing."},{"id":"s47","label":"SQ47","set":"advanced","topicId":"advanced-cost-optimization","text":"A company uses uniform machine types for all workloads regardless of actual CPU/RAM needs. Which optimization is most likely to save cost?","options":["Right-size using custom machine types based on actual utilization","Increase all to the largest machine type","Switch all workloads to local SSD","Move everything to a single zone"],"answer":[0],"why":"Custom machine types allow precise CPU and memory configurations matched to workload needs, eliminating the waste of standard oversized types. Google's Active Assist recommendations help identify right-sizing opportunities."},{"id":"s48","label":"SQ48","set":"advanced","topicId":"advanced-cost-optimization","text":"A company wants to compare the cost impact of Flexible CUDs vs resource-based CUDs. Which statement is correct?","options":["Flexible CUDs apply only to Compute Engine VMs in a specific zone","Resource-based CUDs are scoped to specific machine families and regions; Flexible CUDs can apply across machine families","There is no difference","Flexible CUDs are more expensive per unit than on-demand"],"answer":[1],"why":"Resource-based CUDs are cheaper but locked to a specific machine family and region. Flexible CUDs cost slightly more per unit but provide flexibility to apply across machine families, which is valuable when workloads change over time."},{"id":"s49","label":"SQ49","set":"advanced","topicId":"advanced-cost-optimization","text":"A GKE cluster has many pods requesting 2 vCPU but using only 0.3 vCPU on average. Which optimization reduces wasted compute?","options":["Enable Vertical Pod Autoscaler (VPA)","Increase the resource requests to 4 vCPU","Disable autoscaling entirely","Move all pods to a single node"],"answer":[0],"why":"VPA automatically adjusts pod resource requests based on actual usage, reducing over-provisioning. Increasing requests worsens the waste. Disabling autoscaling removes responsiveness."},{"id":"s50","label":"SQ50","set":"advanced","topicId":"advanced-cost-optimization","text":"A team stores production application logs at full verbosity (DEBUG level) in Cloud Logging and retains them for 365 days. Monthly logging costs are very high. Which combination of actions best reduces cost while keeping useful operational data? (Select TWO)","options":["Reduce production log level to INFO or WARN","Use exclusion filters for low-value log entries","Disable all logging","Store logs in Bigtable instead"],"answer":[0,1],"why":"Reducing log verbosity and excluding known low-value entries (health checks, debug noise) are the safest cost reductions. Disabling all logging creates operational blindness. Bigtable is not designed for log storage."},{"id":"s51","label":"SQ51","set":"advanced","topicId":"advanced-cost-optimization","text":"A company reserves 100 slots for BigQuery but only uses 60 on average, with peaks of 150. Which approach best balances cost and performance?","options":["Keep 100 reserved and add autoscaling for burst capacity","Reserve 200 to cover peaks","Switch entirely to on-demand","Disable reservation and use a spreadsheet"],"answer":[0],"why":"Baseline reservations for steady-state usage combined with autoscaling slots for peak demand provides the best cost-performance balance. Over-reserving wastes money. Pure on-demand is expensive at this scale."},{"id":"s52","label":"SQ52","set":"advanced","topicId":"advanced-cost-optimization","text":"Network egress charges are the highest line item in a company's bill. Most egress is between services in us-central1 and europe-west1. Which architecture change most reduces this cost?","options":["Move inter-region chatty services to the same region and replicate read data where needed","Add Cloud CDN between internal services","Switch to premium network tier","Use larger VM instances"],"answer":[0],"why":"Co-locating chatty services eliminates inter-region egress. Data that must be in multiple regions can use targeted replication, which is cheaper than constant cross-region API traffic. Cloud CDN is for external users, not internal service-to-service."},{"id":"s53","label":"SQ53","set":"advanced","topicId":"advanced-reliability-and-operations","text":"A company needs RPO < 1 minute and RTO < 5 minutes for its transaction database. Which DR pattern is required?","options":["Cold standby with weekly backups","Hot standby with synchronous replication","Warm standby with daily snapshots","Backup to Cloud Storage Coldline"],"answer":[1],"why":"Sub-minute RPO requires synchronous (or near-synchronous) replication, and sub-5-minute RTO requires a hot standby that can serve traffic almost immediately. Cold and warm standby have higher RTO/RPO."},{"id":"s54","label":"SQ54","set":"advanced","topicId":"advanced-reliability-and-operations","text":"Which DR pattern is defined as: \"Full infrastructure is deployed in a secondary region but runs at minimal capacity, scaling up only during failover\"?","options":["Backup and restore","Pilot light","Warm standby","Active-active"],"answer":[2],"why":"Warm standby maintains a running (but scaled-down) duplicate environment that can be scaled up during failover. Pilot light keeps only the most critical components running. Active-active serves traffic from both regions continuously."},{"id":"s55","label":"SQ55","set":"advanced","topicId":"advanced-reliability-and-operations","text":"An SRE team defines an SLO of 99.9% availability for an API. Last quarter, actual availability was 99.95%. The error budget remaining is 50%. What should the team do?","options":["Slow down releases because the error budget is half gone","Proceed with normal release velocity since budget is not exhausted","Stop all changes immediately","Remove monitoring"],"answer":[1],"why":"With 50% error budget remaining and the quarter not yet over, the team has bandwidth for innovation and releases. SRE practice says to release normally when within budget and slow down only when budget is nearly exhausted."},{"id":"s56","label":"SQ56","set":"advanced","topicId":"advanced-reliability-and-operations","text":"A company's Cloud SQL instance fails over to the standby. The application experiences 30 seconds of connection errors. Which improvement most reduces failover impact?","options":["Use connection pooling with retry logic and short timeouts","Increase the Cloud SQL machine type","Remove the failover replica","Switch to manual failover only"],"answer":[0],"why":"Connection pooling with retries and appropriate timeouts allows the application to gracefully handle the brief connection disruption during failover. Increasing size doesn't improve failover speed. Removing the replica eliminates HA."},{"id":"s57","label":"SQ57","set":"advanced","topicId":"advanced-reliability-and-operations","text":"A platform team wants to create log-based metrics that track the count of HTTP 500 errors in application logs and alert when the rate exceeds a threshold. Which workflow is correct?","options":["Create a log-based metric in Cloud Logging → Create an alerting policy in Cloud Monitoring","Export all logs to BigQuery → Run a daily SQL query","Manually grep logs hourly","Disable 500 errors in the application"],"answer":[0],"why":"Log-based metrics extract counter or distribution metrics from log entries, and Cloud Monitoring alerting policies trigger notifications when thresholds are breached—all in near real-time."},{"id":"s58","label":"SQ58","set":"advanced","topicId":"advanced-reliability-and-operations","text":"A multi-region deployment uses Global External Application Load Balancer. Region-A backend becomes unhealthy. What happens to traffic automatically?","options":["Traffic is dropped until manual intervention","The load balancer routes traffic to healthy backends in Region-B","DNS must be manually updated","Users see a 503 error for 30 minutes"],"answer":[1],"why":"The Global External Application Load Balancer continuously health-checks backends and automatically routes traffic away from unhealthy backends to healthy ones in other regions—no manual intervention required."},{"id":"s59","label":"SQ59","set":"advanced","topicId":"advanced-reliability-and-operations","text":"A team wants to implement chaos engineering for their GKE-based microservices. Which approach is most appropriate?","options":["Intentionally inject failures (pod kills, network delays) in non-production and measure SLIs","Delete the production database to test DR","Turn off monitoring during the test","Only test during a real outage"],"answer":[0],"why":"Chaos engineering intentionally introduces controlled failures in a non-production environment (or carefully in production with safeguards) while observing SLI impact. Deleting production databases is destructive, not controlled experimentation."},{"id":"s60","label":"SQ60","set":"advanced","topicId":"advanced-reliability-and-operations","text":"Cloud Profiler reveals that a specific microservice has high CPU consumption from a particular function. What is the recommended next step?","options":["Increase the CPU allocation for all services","Investigate and optimize the specific hot function identified by the profiler","Disable Cloud Profiler","Add more regions"],"answer":[1],"why":"Cloud Profiler identifies specific code-level hot spots. The right action is to investigate and optimize the identified function, not blindly add resources across all services."},{"id":"s61","label":"SQ61","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A company completing a phased migration wants to manage DNS records that resolve to either on-premises or Google Cloud services depending on migration progress. Which service supports this?","options":["Cloud DNS with split-horizon configuration","Hard-code IP addresses in application config","Use Cloud CDN as DNS resolver","Create separate domain names for each service"],"answer":[0],"why":"Cloud DNS split-horizon (private + public zones) allows the same DNS names to resolve differently based on the query source—on-prem resolvers can point to on-prem, while GCP networks resolve to cloud backends—facilitating gradual migration."},{"id":"s62","label":"SQ62","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A company wants to move its on-premises Hadoop ecosystem to Google Cloud. They plan to keep the same Spark jobs and HDFS storage patterns initially. Which is the best migration target?","options":["Dataproc with Cloud Storage as the storage layer","BigQuery only","Cloud Run","Memorystore"],"answer":[0],"why":"Dataproc is the managed Hadoop/Spark service. Using Cloud Storage (via the HDFS connector) as the storage layer replaces HDFS with a more scalable, durable, and cost-effective solution—with minimal job rewrites."},{"id":"s63","label":"SQ63","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A multinational company runs workloads across AWS, Azure, and Google Cloud. They need a single place to manage security policies and see compliance posture across all three. Which Google Cloud product addresses this?","options":["Security Command Center Enterprise (multi-cloud)","Cloud Armor only","VPC peering","Artifact Registry"],"answer":[0],"why":"Security Command Center Enterprise (premium tier) provides multi-cloud security posture management with visibility into findings across Google Cloud, AWS, and Azure."},{"id":"s64","label":"SQ64","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A company is migrating VMs but wants to assess which workloads should be lifted-and-shifted vs modernized. Which tool helps with this assessment?","options":["Migration Center (formerly StratoZone)","Cloud CDN analytics","Cloud Scheduler","BigQuery BI Engine"],"answer":[0],"why":"Migration Center provides discovery, assessment, and planning tools that analyze on-premises workloads and recommend migration strategies (rehost, replatform, refactor)."},{"id":"s65","label":"SQ65","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A company needs to replicate VMware VMs to Google Cloud for DR with continuous replication. Which service is designed for this?","options":["Migrate to Virtual Machines (M2VM)","Storage Transfer Service","Transfer Appliance","Datastream"],"answer":[0],"why":"Migrate to Virtual Machines supports continuous replication of VMware VMs to Compute Engine, enabling DR failover. Storage Transfer Service handles object storage, not VMs. Datastream is for database CDC."},{"id":"s66","label":"SQ66","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A team migrating databases uses Database Migration Service and encounters schema incompatibilities between source MySQL 5.7 and target Cloud SQL MySQL 8.0. What is the recommended approach?","options":["Fix schema incompatibilities before migration and test with a dry-run migration","Force migration and fix issues in production","Move to Bigtable instead","Ignore errors and switch DNS immediately"],"answer":[0],"why":"Pre-migration testing and schema compatibility fixes are essential. Forcing migration with known issues risks data loss. Always validate with a dry run."},{"id":"s67","label":"SQ67","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"A company needs to migrate a SAP HANA workload to Google Cloud with certified configurations. Which specialization should they look for?","options":["Bare Metal Solution or certified Compute Engine configurations for SAP","Move SAP to Cloud Functions","Run SAP on Cloud Run","Use Firestore instead of HANA"],"answer":[0],"why":"Google Cloud provides SAP-certified machine configurations on Compute Engine and Bare Metal Solution for larger HANA instances. SAP workloads cannot run on serverless platforms."},{"id":"s68","label":"SQ68","set":"advanced","topicId":"hybrid-multi-cloud-and-migration","text":"An enterprise wants to co-locate private Google Cloud access with their on-premises infrastructure using a Google-operated facility nearby. Which setup supports this model?","options":["Cross-Cloud Interconnect","Cloud CDN signed URLs","Public peering only","Cloud Shell"],"answer":[0],"why":"Cross-Cloud Interconnect provides dedicated connectivity between Google Cloud and other cloud providers or on-premises facilities, with high bandwidth and low latency through Google-operated infrastructure."},{"id":"s69","label":"SQ69","set":"advanced","topicId":"bigquery-advanced-and-data-analytics","text":"A BI team needs sub-second query responses on a frequently accessed BigQuery dashboard dataset of 200 GB. Which feature helps most?","options":["BigQuery BI Engine","Export to Memorystore","Use Cloud SQL instead","Add more partitions"],"answer":[0],"why":"BI Engine provides in-memory analysis for sub-second interactive query responses on dashboards. It's designed exactly for accelerating BI workloads on smaller hot datasets."},{"id":"s70","label":"SQ70","set":"advanced","topicId":"bigquery-advanced-and-data-analytics","text":"A data governance team needs to restrict access so that analysts in the EU can only see rows belonging to EU customers in a BigQuery table. Which feature enables this?","options":["Row-level security policies in BigQuery","Column-level encryption only","Create separate tables per region","Use Cloud DLP to mask the table"],"answer":[0],"why":"BigQuery row-level security policies filter rows based on the querying user's attributes, enabling data segmentation without duplicating tables. Column-level security hides columns, not rows."},{"id":"s71","label":"SQ71","set":"advanced","topicId":"bigquery-advanced-and-data-analytics","text":"A company wants to expose a BigQuery dataset to a partner for reading but ensure they can only see non-PII columns. Which combination is most appropriate?","options":["Authorized view with column-level security and data masking","Export to CSV and email it","Share the service account key","Create a public dataset"],"answer":[0],"why":"Authorized views combined with column-level security and policy tags that mask PII columns provide granular data sharing without exposing sensitive fields."},{"id":"s72","label":"SQ72","set":"advanced","topicId":"bigquery-advanced-and-data-analytics","text":"A streaming application posts 100K events/second into BigQuery. Which ingestion method is most appropriate?","options":["BigQuery Storage Write API (streaming)","Batch load from Cloud Storage daily","INSERT DML statements from application code","gsutil cp from local disk"],"answer":[0],"why":"The BigQuery Storage Write API provides high-throughput, low-latency streaming ingestion for large event volumes. Batch loading introduces delay. DML inserts don't scale to 100K events/second."},{"id":"s73","label":"SQ73","set":"advanced","topicId":"bigquery-advanced-and-data-analytics","text":"A data team wants to pre-compute expensive aggregations that auto-refresh when base table data changes. Which BigQuery feature supports this?","options":["Materialized views","Standard views only","External tables","Federated queries"],"answer":[0],"why":"Materialized views store pre-computed results and automatically refresh incrementally when the base table changes, reducing query cost and latency for repeated aggregation patterns."},{"id":"s74","label":"SQ74","set":"advanced","topicId":"bigquery-advanced-and-data-analytics","text":"A company uses Looker for dashboards and wants to ensure consistent metric definitions across teams. Which Looker feature provides this?","options":["LookML data modeling layer","Ad-hoc SQL in each dashboard","Export CSVs to each team","Duplicate dashboards per department"],"answer":[0],"why":"LookML provides a centralized semantic model layer where business logic, metrics, and dimensions are defined once and reused across all dashboards—ensuring consistency across teams."},{"id":"s75","label":"SQ75","set":"advanced","topicId":"well-architected-framework-and-architecture-patterns","text":"The Google Cloud Well-Architected Framework includes which pillars? (Select TWO that are correct)","options":["Operational Excellence and Reliability","Sustainability and Cost Optimization","Marketing and Brand Awareness","Color Theory and UX Design"],"answer":[0,1],"why":"The Well-Architected Framework pillars are: Operational Excellence, Security, Reliability, Performance Optimization, Cost Optimization, and Sustainability. Marketing and UX design are not cloud architecture pillars."},{"id":"s76","label":"SQ76","set":"advanced","topicId":"well-architected-framework-and-architecture-patterns","text":"A global e-commerce company wants its architecture to survive the failure of an entire Google Cloud region with zero data loss. Which architecture pattern is required?","options":["Multi-region active-active with synchronous replication","Single-zone deployment with daily backups","Warm standby with asynchronous replication (RPO > 0)","Cold DR with weekly exports"],"answer":[0],"why":"Zero data loss (RPO = 0) during a regional failure requires synchronous replication and an active-active deployment. Asynchronous replication has RPO > 0 (some data loss). Cold DR has the highest RPO."},{"id":"s77","label":"SQ77","set":"advanced","topicId":"well-architected-framework-and-architecture-patterns","text":"A microservices architecture uses the \"strangler fig\" pattern. What does this mean?","options":["Build a new monolith to replace all services at once","Gradually route traffic from legacy monolith functions to new microservices until the monolith is fully replaced","Delete the old system immediately","Run both old and new systems permanently without migration"],"answer":[1],"why":"The strangler fig pattern incrementally replaces monolith functionality by routing requests to new services one feature at a time, reducing migration risk."},{"id":"s78","label":"SQ78","set":"advanced","topicId":"well-architected-framework-and-architecture-patterns","text":"An architect must design a system where a service failure does not cascade to dependent services. Which combination of patterns is most effective? (Select TWO)","options":["Circuit breaker pattern","Retry with exponential backoff and jitter","Infinite immediate retries","Tight synchronous coupling"],"answer":[0,1],"why":"Circuit breakers stop calling a failing service after a threshold, preventing cascade. Exponential backoff with jitter reduces thundering herd effects during recovery. Infinite immediate retries worsen overload. Tight coupling increases cascade risk."},{"id":"s79","label":"SQ79","set":"advanced","topicId":"well-architected-framework-and-architecture-patterns","text":"A fintech application must process exactly-once semantics for financial transactions in an event-driven system. What is the most practical approach on Google Cloud?","options":["Use Pub/Sub with idempotent consumers that deduplicate on a unique transaction ID","Call services synchronously with no retries","Use Memorystore as the primary transaction database","Disable retries to avoid duplicates"],"answer":[0],"why":"True exactly-once is achieved in distributed systems by combining at-least-once delivery (Pub/Sub) with idempotent consumer logic that deduplicates using a unique key. Disabling retries trades correctness for data loss."},{"id":"s80","label":"SQ80","set":"advanced","topicId":"well-architected-framework-and-architecture-patterns","text":"An architect is reviewing a system where developers grant themselves the \"Editor\" primitive role on production projects. What is the main risk and recommended action?","options":["No risk; primitive roles are fine for production","Excessive permissions—migrate to predefined or custom roles with least privilege","Insufficient permissions—grant Owner instead","Primitive roles are not supported anymore"],"answer":[1],"why":"The Editor role grants broad write access across all services in a project, violating least privilege. Predefined roles (or custom roles) should be used to restrict access to only what's needed per workload."}]}