Mavi-Yeşil ve Kanarya Dağıtımları

Replacing all running instances at once is risky: a bad release takes down everyone immediately. Safer strategies roll out changes gradually or keep a tested fallback ready.

Blue-green runs two identical environments. Blue serves live traffic while you deploy and test the new version on green. When green is verified, you switch all traffic to it.

A load balancer or router points to the active color. Switching is just changing where the router sends traffic, which is instant and easy to reverse.

If green misbehaves after the switch, you flip the router back to blue. Because blue is still running the old version, rollback is immediate with no rebuild needed.

A canary release sends a small percentage of traffic to the new version first, e.g. 5%. You watch metrics, then gradually increase to 100% if everything looks healthy.

Route a fraction of users to the new version by weighting at the load balancer or by user attributes. Start small so any failure affects only a few users.

During a canary, monitor error rate, latency, and key business metrics for the canary group versus the baseline. A spike in errors is your signal to stop.

CI/CD tools can auto-promote a canary: bump the weight in steps (5% to 25% to 50% to 100%) only while metrics stay within thresholds, otherwise roll back automatically.

Both strategies assume two versions may run at once. Make schema migrations backward compatible: add columns before using them, and avoid destructive changes until the old version is gone.

Ensure users are not broken by switching versions mid-session. Keep sessions in a shared store and design APIs so old and new versions can coexist during the rollout window.

Use blue-green when you want a clean, instant switch with easy rollback. Use canary when you want to limit blast radius and validate with real production traffic gradually. Many teams combine both.

Check your deployment strategy knowledge.

You learned two safe release strategies:

• Blue-green: two environments with an instant, reversible traffic switch
• Canary: gradual rollout to a small traffic slice with metric-driven promotion
• Keep migrations backward compatible and sessions shared

Both enable zero-downtime deploys with quick recovery from bad releases.