Observability & Reliability
OpenTelemetry, logs/metrics/traces, SLOs with error budgets, chaos drills.
Capabilities
- Contracts & change safety
- SLIs/SLOs
- Cost guardrails
Perennial patterns
- Standards-first
- Separation of concerns
- Failure as a normal state
Modern enablers
- Examples you can swap as needed
Anti-patterns
- Vendor-led architecture without contracts
- Shared-DB coupling
What good looks like
- Green CI gates
- Trace IDs end-to-end
- DR drill in last quarter
Point A → Point B, Fast
Architecture Sprint (1–2 weeks): Current → Target → 1–90-day plan.
Book 30-min intro