deployment-patterns

Deployment Patterns

When to Apply

• Deploying an application for the first time
• Choosing a hosting strategy (serverless, containers, VMs, edge)
• Setting up CI/CD pipelines
• Managing multiple environments (dev, staging, production)
• Implementing zero-downtime deployments
• Preparing an application for production

Rules

CRITICAL (rules/hosting-and-containers.md)

• DEP-01: Hosting strategy selection -- choose based on workload (serverless, containers, PaaS, edge, VMs)
• DEP-02: Containerize every deployable service with multi-stage Dockerfiles
• DEP-03: Three environments minimum (dev, staging, prod); same image artifact, different config only

HIGH (rules/cicd-and-deploys.md)

• DEP-04: CI/CD pipeline: Push -> Lint -> Test -> Build -> Scan -> Deploy Staging -> Smoke -> Deploy Prod
• DEP-05: Zero-downtime deploys via rolling update, blue-green, or canary strategies

HIGH (rules/production-readiness.md)

• DEP-06: Production readiness checklist -- health checks, graceful shutdown, resource limits, structured logging, backups, security headers, alerting
• DEP-07: Rollback strategy -- feature flags first, then code, then database (last resort)

Anti-Patterns

• "Works on my machine" deploys -- no Dockerfile, no CI, manual scp to servers
• Snowflake servers -- configured manually, not reproducible; use IaC
• Secrets in git -- even in private repos; use a secrets manager
• No health checks -- platform cannot tell if app is alive or dead
• Big bang deploys -- all traffic switches at once with no canary or staged rollout
• Coupled database + code deploys -- separate schema changes from code changes
• No rollback plan -- "we'll fix forward" works until it doesn't

Priority Summary

ID	Rule	Priority
DEP-01	Hosting strategy selection	CRITICAL
DEP-02	Containerization	CRITICAL
DEP-03	Environment management	CRITICAL
DEP-04	CI/CD pipeline	HIGH
DEP-05	Zero-downtime deployments	HIGH
DEP-06	Production readiness checklist	HIGH
DEP-07	Rollback strategy	MEDIUM