advanced

Google Cloud / GCP

Google Cloud services for compute, containers, serverless workloads, storage, relational data, messaging, identity, and logging.

GCP interviews emphasize projects, service accounts, Cloud Run for containerized HTTP, GKE when Kubernetes is required, Cloud Storage, Cloud SQL, Pub/Sub, and Cloud Logging. JavaScript teams often compare Cloud Run to App Engine or GKE.

Subtopics: Compute Engine, Cloud Run, GKE, Cloud Functions, Cloud Storage, Cloud SQL, Pub/Sub, IAM, Cloud Logging.

On interviews: deploy a stateless Node container to Cloud Run with a Cloud SQL connection via connector, Pub/Sub for background work, and workload identity instead of JSON keys.

Common pitfalls: downloading service account keys into repos; Pub/Sub consumers without acknowledgment deadlines; Cloud SQL opened to `0.0.0.0/0`.

The trade-off is strong managed Kubernetes and data services versus a smaller default hiring familiarity than AWS.

Checklist:

Use workload identity and scoped service accounts.
Prefer Cloud Run for request-scaled HTTP containers.
Design Pub/Sub with DLQ and idempotent handlers.
Centralize logs with trace correlation.