GCP GCP-PCA Free Practice Questions — Page 1

Correct Answer: D. Use separate backend pools for each API path behind the load balancer

Using separate backend pools for each API version behind a single load balancer is the correct approach. Google Cloud HTTP(S) Load Balancer supports URL map rules that route traffic to different backend services based on the request path (e.g., /v1/* → backend-v1, /v2/* → backend-v2). This keeps the same SSL certificate and DNS records intact, satisfying that requirement. Option A is wrong because adding a new load balancer would require new DNS/SSL configuration. Option B forces old clients to change, which contradicts keeping the old API available transparently. Option C adds unnecessary complexity and coupling between API versions — the old API shouldn't need to know about the new one. Option D cleanly separates concerns at the infrastructure level without touching application code.

Correct Answer: A. Load data into Google BigQuery

BigQuery is Google Cloud's fully managed, serverless data warehouse designed for petabyte-scale analytics using standard SQL. It is the best fit because analysts already know SQL and no new tooling is required. BigQuery supports 24/7 availability with built-in redundancy and high durability. Cloud SQL (Option B) is a relational database but is not designed for multi-petabyte datasets — it has storage limits and is optimized for transactional workloads. Cloud Storage (Option C) stores flat files but does not provide a native SQL query interface for analysts. Cloud Datastore (Option D) is a NoSQL document database, not suitable for SQL-based analytics or petabyte-scale structured data.

Correct Answers: C. Instrument the application with a monitoring tool like Stackdriver Debugger; D. Select an automation framework to reliably provision the cloud infrastructure; E. Deploy a continuous integration tool with automated testing in a staging environment

These three practices represent cloud migration best practices for enterprise applications. Option C — instrumenting with Stackdriver (now Cloud Monitoring/Debugger) — is essential for observability during and after migration. Option D — using an infrastructure automation framework (like Terraform or Deployment Manager) — ensures repeatable, reliable provisioning and reduces human error. Option E — setting up CI/CD with automated testing in staging — validates that the migrated application behaves correctly before production. Option A is incorrect because rewriting a J2EE app for App Engine is a full re-platform, not a migration practice recommendation. Option B is incorrect because Cloud Dataflow is a data pipeline tool, unrelated to capturing app metrics. Option F is incorrect because migrating from MySQL to NoSQL is an unnecessary architectural change that adds risk and complexity without clear benefit during migration.

Correct Answer: A. The session variable is local to just a single instance

The sessions variable is a Python dictionary defined at the module level (sessions = {}). In App Engine, multiple instances run simultaneously during peak load. Each instance has its own in-memory copy of sessions — they are not shared between instances. So when a user's request is routed to a different instance than their previous request, that instance has no record of which articles the user already viewed, causing previously viewed articles to appear again. The fix is to store session state in a shared external store like Cloud Datastore, Memorystore (Redis), or Cloud Firestore — accessible by all instances. Option B is wrong — the code never writes to Cloud Datastore. Option C and D are wrong — this is a server-side state problem, not a client-side caching issue.

Correct Answer: C. Help them define their requirements and assess viable logging tools

Before recommending or implementing any tool, the correct engineering approach is to define requirements first, then evaluate options. Option A jumps to a solution (Stackdriver) without understanding their needs. Option B is too passive and unhelpful. Option D assumes the current tool can be fixed, contradicting the team's assessment. A structured requirements-and-assessment process ensures the chosen tool actually solves their problem and avoids costly rework.