Canary release
Also called canary deployment, progressive rollout.
A release method that sends a small share of users to a new version first, watches for problems, and widens the rollout step by step.
Last updated
What it is.
The name comes from the canaries miners took underground as an early warning. A canary release might start at one percent of traffic, then five, then twenty-five, then everyone, while error rates and response times are compared with the old version.
If the numbers get worse, traffic goes back to the old version and only a few users were affected. Canary releases can be done at the router level or in code with a feature flag.
Why it matters
Some bugs only show up with real users and real data. A canary limits how many people meet them. It is especially useful for high-traffic products where even a few minutes of a bad release would hurt.
In a product build
We automate the steps and the checks, so the rollout pauses itself when error rate or latency crosses a threshold. Each step lasts long enough to see normal traffic, and dashboards compare canary and stable side by side.
Common mistakes.
- 01
Rolling out faster than the metrics can show a problem.
- 02
No automatic stop, so a bad canary only gets noticed by support tickets.
- 03
Canary groups that are not typical, such as only internal staff.
Related terms.
- Blue-green deploymentA release method that runs two identical production environments and switches all traffic from the old one to the new one in one step.
- Feature flagA switch in the code that turns a feature on or off for chosen users at runtime, without deploying new code.
- ObservabilityHow well you can understand what a running system is doing, and why, from the data it gives off: logs, metrics and traces.
- CI/CD (continuous integration and delivery)An automated pipeline that builds and tests every code change (CI) and then releases it to users in a repeatable way (CD).
- A/B testAn experiment that shows two versions of something to random groups of users and measures which one does better on a chosen metric.
Guides and pages
Browse the full glossary or read answers to common questions.
