fluxcd-flagger

What is fluxcd-flagger?

The Flagger image packages Flagger, an open-source progressive delivery operator maintained by FluxCD. It automates the release process for Kubernetes workloads by gradually shifting traffic between versions while monitoring metrics, logs, and health checks.

Flagger supports deployment strategies such as canary releases, blue/green deployments, and A/B testing. It integrates with service meshes and ingress controllers including Istio, Linkerd, App Mesh, NGINX, and Gloo, allowing fine-grained traffic control during rollouts.

By automating promotion and rollback decisions based on observability data from providers like Prometheus and Datadog, Flagger helps teams reduce risk, shorten feedback loops, and improve reliability when shipping new versions of applications.

How to use this image

The Flagger image is deployed as a controller in a Kubernetes cluster.

Examples:

<code># Install Flagger with Helm
helm repo add flagger <https://flagger.app>
helm upgrade -i flagger flagger/flagger \\
 --namespace=flagger-system \\
 --set meshProvider=linkerd \\
 --set metricsServer=http://prometheus:9090</code>

<code># Apply a canary deployment
kubectl apply -f canary.yaml</code>

<code># Monitor the canary rollout
kubectl describe canary podinfo</code>

Canary custom resources (CRDs) define the rollout strategy, traffic routing, and metrics analysis rules that Flagger uses to automate progressive delivery.

Image variants

The Flagger image is generally published under ghcr.io/fluxcd/flagger and available in these forms:

flagger:<version>

Version-pinned images (e.g., flagger:1.33.0) are recommended for production use to ensure consistency.

flagger:latest

Tracks the most recent stable release. Useful for evaluation but not ideal for production due to version drift.

flagger:canary

Built from the development branch with the latest commits. Intended for contributors or teams testing new features.