Architecture Patterns
Proven structural solutions to recurring architectural problems. Patterns are not implementations — they’re descriptions of proven approaches with trade-off analysis.
What’s Here
- architecture-patterns — Catalog of cloud-native architecture patterns
Common Patterns Quick Reference
| Pattern | What It Solves | Example |
|---|---|---|
| Strangler Fig | Incremental migration from legacy | Route traffic to new system piece by piece |
| Sidecar | Attach utilities to services without modifying them | Logging sidecar, metrics exporter |
| Circuit Breaker | Prevent cascading failures | Hystrix, Envoy circuit breaker |
| CQRS | Separate read and write models | Event-sourced systems |
| Event Sourcing | Store state changes as events | Audit trails, temporal queries |
| Saga | Distributed transactions without2PC | Choreography vs orchestration |
| Bulkhead | Isolate failures | Separate thread pools per dependency |
| Leader Election | Single-writer coordination | etcd, Zookeeper |
| Write-Ahead Log | Durability before acknowledgment | Kafka, PostgreSQL WAL |
When to Use Patterns
Patterns are starting points, not finished designs. Every pattern introduces trade-offs:
Strangler Fig:
✅ Incremental migration, low risk
✅ Old and new systems coexist
❌ Dual-running complexity
❌ Temporary twice the infrastructure cost
Related
- thinking-like-an-architect — How to evaluate pattern trade-offs
- resilience — Implementation of resilience patterns
- README — System design interview patterns