Flagger on Kubernetes: Progressive Delivery and Canary Deployments
Deploying fast is easy; deploying safely is hard. Instead of swapping Kubernetes resources by hand or hoping a rollout won’t break users, Flagger runs the release process itself: it creates a safe duplicate of your deployment, controls how traffic shifts between versions, and watches live signals so unhealthy builds get pulled back automatically.
In this episode, Whitney Lee and Stefan Prodan, maintainer of Flux and Flagger, show how progressive delivery works in real Kubernetes clusters.
They walk through canary rollouts step by step and explain why GitOps needs that safe duplicate. But how does Flagger check health? They answer: first with built-in error and latency signals, then with custom checks from systems like Prometheus or Datadog, and finally with optional hooks into your own tools.
They also compare different rollout styles: blue/green, blue/green with mirrored traffic, A/B testing based on headers or cookies, and canary with “sticky sessions” that keep each user on the same version. The result is release control that fits smoothly with Flux, Argo, or plain CI-to-Helm workflows.
Learn how Flagger automates progressive delivery in Kubernetes, making it safe to deploy on Fridays. This powerful tool handles canary deployments, blue-green deployments, and A/B testing while you enjoy your weekend. With automatic rollbacks and metric analysis, Flagger ensures your deployments stay healthy even when you're away.
Prefer the full ⚡ Enlightning episode?
Fearless Delivery with Flagger → https://www.youtube.com/live/U6tG6n7XYFI
Watch all ⚡ Enlightning episodes → • ⚡️ Enlightning
Watch all 🌩️ Thunder episodes → • 🌩️ Thunder
٩( ᐛ )و Subscribe to Whitney’s YouTube channel → / @wiggitywhitney
#Flagger #Kubernetes #ProgressiveDelivery #Flux #CanaryDeployments
▬▬▬▬▬▬ ⚡️ Related Links ⚡️ ▬▬▬▬▬▬
🔗 Flagger → https://fluxcd.io/flagger/
🔗 Istio → https://istio.io/
🔗 CNCF Slack → https://slack.cncf.io/
🔗 Flux → https://fluxcd.io/