Overview of Canary Analysis using Kayenta for Spinnaker pipelines

Kayenta is the Spinnaker microservice that is responsible for enabling canary deployments in Spinnaker and performing automated canary analysis. Canary deployment is a release deployment strategy where a new release version is deployed alongside the stable running version.

It is exposed to a small subset of the end-users and then evaluated against the baseline (a copy of current deployment) to find out if the new release is stable and ready to be rolled out. The application config in Spinnaker has the configuration option to activate canary config and canary reports on Deck.

Kayenta supports monitoring tools like Stackdriver, Prometheus, Datadog, New Relic, e.t.c. Performing automated canary analysis with Kayenta involves two stages — Metric Retrieval and Judgement. Key metrics are retrieved from canary and baseline clusters, tagged, and stored in a time-series database.

This data is then validated, cleaned, compared and a score is then computed for each metric. An overall high score results in a favorable judgment for the canary or else it is failed. Please read the full blog “Overview of Canary Analysis using Kayenta for Spinnaker pipelines” to know more about the configurations and best practices for automated canary analysis using Kayenta.