deploy-strategy

Deployment Strategy

Plan and document a deployment strategy for a service.

Arguments

Strategy and service: $ARGUMENTS

Workflow

1. Assess the service

   • If a service name or path is provided, analyze the codebase
   • Identify: stateful vs stateless, database dependencies, external integrations
   • Check for existing deployment configs (Dockerfile, k8s manifests, terraform, docker-compose)

2. Recommend strategy (if not specified)

   • Rolling: Default for most stateless services. Zero-downtime, gradual replacement.
   • Blue-Green: When you need instant rollback. Two identical environments.
   • Canary: When you need to validate with real traffic. Gradual traffic shift.
   • Decision factors: risk tolerance, infrastructure cost, rollback speed, traffic patterns

3. Design the strategy

   Rolling:

   • Max unavailable / max surge settings
   • Health check configuration (readiness + liveness probes)
   • Graceful shutdown handling (drain connections before termination)

   Blue-Green:

   • Environment provisioning (identical blue and green)
   • Traffic switch mechanism (DNS, load balancer, router)
   • Database migration strategy (backward-compatible schemas)
   • Smoke test suite to run before switching

   Canary:

   • Traffic split percentages and timeline (1% → 5% → 25% → 100%)
   • Success criteria (error rate, latency percentiles, business metrics)
   • Automatic rollback triggers
   • Monitoring dashboard requirements

4. Generate artifacts

   • Deployment configuration files (k8s, terraform, or platform-specific)
   • Runbook: step-by-step deployment procedure
   • Rollback procedure
   • Monitoring checklist

5. Output

   • Strategy recommendation with rationale
   • Configuration files
   • Deployment runbook
   • Rollback plan
   • Monitoring and alerting requirements