Skip to content
LiveNew tools this week
Submit
AIMarketly
gcp-patterns logo

GCP Production Patterns

gcp-patterns

GCP production patterns: Cloud Run with Terraform, Workload Identity Federation (no service account keys), Cloud SQL with private IP, Memorystore Redis, BigQuery for analytics, Cloud Armor WAF, Secret Manager, VPC Service Controls, IAM least privilege bindings. Use when designing GCP architecture...

rnavarych/alpha-engineer0installs15starsDevOps & CI/CD

SKILL.md

Full skill instructions

GCP Production Patterns

When to Use This Skill

  • Deploying containers to Cloud Run
  • Setting up Workload Identity Federation (no SA keys)
  • Configuring Cloud SQL with private IP
  • Writing Terraform for GCP infrastructure
  • Querying BigQuery for analytics

Core Principles

  1. Cloud Run for stateless services — autoscales to 0, no cluster management, pay per request
  2. Workload Identity Federation, never SA keys — SA keys are persistent credentials; WIF is token-based and short-lived
  3. Private IP for all data services — Cloud SQL, Memorystore: no public IP, accessed via VPC
  4. IAM bindings on resources, not members — bind roles to service accounts scoped to specific resources
  5. BigQuery for analytics over Cloud SQL — even 1B row aggregations run in <10 seconds on BigQuery

References available

  • references/​compute-patterns.md — Cloud Run Terraform, Workload Identity Federation for GitHub Actions, IAM least privilege bindings
  • references/​data-patterns.md — Cloud SQL private IP, Memorystore Redis, BigQuery partitioning and clustering, Secret Manager
  • references/​networking-patterns.md — VPC connector, Cloud Armor WAF, VPC Service Controls, load balancer setup