Release Engineering¶

Most incidents follow a change, so safe releases are one of the best reliability investments.

Principles¶

Small, frequent changes are easier to understand and roll back.
Automate the path to production, and review everything through GitOps where you can.
Release progressively, and watch the SLIs while you do.
Make rollback fast and boring.

Strategies¶

Strategy	How it works
Rolling update	Replace Pods gradually (the Kubernetes default).
Canary	Send a small share of traffic to the new version first.
Blue/green	Run two environments and switch traffic.
Feature flags	Ship code dark and enable it separately.

Safe rolling updates¶

strategy:
  type: RollingUpdate
  rollingUpdate:
    maxUnavailable: 0
    maxSurge: 25%

kubectl rollout status deploy/api
kubectl rollout undo deploy/api

Measure delivery¶

The DORA metrics are a good scorecard: deployment frequency, lead time for changes, change failure rate, and time to recover from a failed deployment.

Tie release speed to the error budget: when the budget is healthy, release more freely.