Problem
Deployments cause outages. Big bang = 100% blast radius. Need progressive delivery with automated rollback.
deploymentestablished · high operational burden
Deployment Strategies
Also known as: deployment-patterns, blue-green, canary, rolling, feature-flags, dark-launch
Intent
Choose and implement safe deployment strategies: rolling, blue-green, canary, feature flags — with automated rollback and observability.
Forces
- •Zero downtime required
- •Instant rollback capability
- •Real traffic validation before full rollout
- •Automation: no manual approval gates
Solution
✓ When to Use
- •Any production deployment
- •Zero-downtime requirements
- •Risk mitigation for changes
✗ When Not to Use
- •Dev/test environments (rolling fine)
- •Single-instance services (no HA)
- •Team not ready for automation investment
Pros
- +Zero downtime deployments
- +Instant rollback capability
- +Risk mitigation: blast radius control
- +Deploy frequency decoupled from release
Cons
- −Infrastructure overhead (blue-green 2x, canary 2 versions)
- −Automation complexity (metrics, analysis, rollback)
- −Schema changes: expand/contract discipline
- −Flag debt: cleanup process required
Cost Profile
Infrastructure
Low (rolling) to High (blue-green 2x)
Operational
High — automation, metrics, runbooks
Cognitive
Medium — strategy choice, automation design
Failure Modes
Canary metrics wrong: false promote/rollback
Flag debt: 200 flags, no cleanup
Schema break: expand/contract violated
Rollback fails: traffic switch broken, manual intervention
Dark launch cost: 2x compute, no user signal
Real-World Examples
Alternatives
- rolling-update
- recreate
- big-bang
- feature-flags
- dark-launch
Related Patterns
- feature-flag
- canary-deployment
- blue-green-deployment
- dark-launch
- expand-contract
- database-migration
- observability
- slo-slo
Competency Domains
deploymentreliability opseconomics evolutiondistribution communicationdata state