Deployment Skill

=== PURPOSE ===

This skill provides procedures for deployment automation and execution. Used by the DevOps Engineer in coordination with Delivery Lead.

=== USED BY ===

=== PREREQUISITES ===

Before using this skill:

• Infrastructure provisioned
• CI/CD pipeline ready
• Release approved by Delivery Lead

=== PROCEDURE: Deployment Checklist ===

Pre-Deployment:

• All tests passing
• Build artifacts ready
• Database migrations prepared
• Environment variables configured
• Rollback plan documented
• Team notified

Deployment:

• Take backup (if applicable)
• Run database migrations
• Deploy application
• Run health checks
• Verify functionality

Post-Deployment:

• Monitor logs
• Check error rates
• Verify key flows
• Update status
• Log in buildlog

=== PROCEDURE: Zero-Downtime Deployment ===

Strategies:

Blue-Green:

1. Deploy to inactive environment (green)
2. Run tests on green
3. Switch traffic to green
4. Keep blue for rollback

Rolling:

1. Deploy to subset of servers
2. Verify health
3. Continue to next subset
4. Complete when all updated

Canary:

1. Deploy to small percentage
2. Monitor metrics
3. Gradually increase percentage
4. Full rollout when confident

=== PROCEDURE: Database Migration ===

Steps:

1. Backup database
2. Run migrations in staging first
3. Apply to production during low-traffic
4. Verify data integrity
5. Keep rollback scripts ready

Rules:

• Migrations must be backwards compatible
• Test rollback scripts
• Never drop columns immediately

=== PROCEDURE: Rollback ===

When to Rollback:

• Critical errors in production
• Performance degradation
• Data integrity issues
• Security vulnerabilities

Steps:

1. Notify team immediately
2. Switch to previous version/environment
3. Rollback database if needed
4. Verify functionality
5. Document incident
6. Log with #issue-encountered

=== PROCEDURE: Health Checks ===

Endpoint: /health or /healthz

Response:

{
  "status": "healthy",
  "version": "1.2.3",
  "timestamp": "2026-01-04T12:00:00Z",
  "checks": {
    "database": "healthy",
    "cache": "healthy",
    "external_api": "healthy"
  }
}

Use Cases:

• Load balancer health checks
• Kubernetes liveness/readiness probes
• Monitoring systems

=== PROCEDURE: Deployment Documentation ===

Runbook Contents:

• Step-by-step deployment process
• Environment-specific configurations
• Rollback procedures
• Contact information
• Common issues and solutions

Location: docs/deployment-runbook.md

=== ANTI-PATTERNS ===

No Rollback Plan

Problem: Can't recover from bad deployment Solution: Always have rollback ready

Manual Deployments

Problem: Error-prone, inconsistent Solution: Automate everything

No Health Checks

Problem: Can't verify deployment success Solution: Implement health endpoints

Big Bang Deployments

Problem: High risk, hard to debug Solution: Use incremental strategies

=== RELATED SKILLS ===