Argo Rollouts

Configuring Rollout resources, AnalysisTemplates, Prometheus metrics, pause steps, and traffic routers.

8 min readCI/CD & GitOpsPlatform

Rollout Resource

Argo Rollouts extends Kubernetes with a Rollout CRD that replaces the standard Deployment. It supports canary and blue-green strategies with automated analysis and traffic shaping. The Rollout resource is a drop-in replacement for Deployment — it uses the same pod template spec.

Basic Rollout resource


apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: myapp
spec:
  replicas: 5
  revisionHistoryLimit: 5
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
        - name: myapp
          image: ghcr.io/myorg/myapp:2.0.0
          ports:
            - containerPort: 8080
          readinessProbe:
            httpGet:
              path: /health
              port: 8080
            initialDelaySeconds: 10
            periodSeconds: 5

The Rollout resource looks identical to a Deployment. The difference is in the strategy section. Argo Rollouts controller watches Rollout resources and manages the progressive delivery lifecycle.

Canary Strategy

Canary rollout with steps


apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: myapp
spec:
  replicas: 10
  strategy:
    canary:
      canaryService: myapp-canary
      stableService: myapp-stable
      trafficRouting:
        nginx:
          stableIngress: myapp-ingress
      steps:
        - setWeight: 10
        - pause: { duration: 5m }
        - setWeight: 25
        - pause: { duration: 5m }
        - setWeight: 50
        - pause: { duration: 10m }
        - setWeight: 75
        - pause: { duration: 5m }
        - setWeight: 100
      analysis:
        templates:
          - templateName: canary-analysis
        startingStep: 2
        args:
          - name: service-name
            value: myapp-canary

The canary strategy gradually increases traffic weight. At each step, AnalysisTemplates verify the canary is healthy. If analysis fails, the rollout is aborted. The stableService receives traffic for non-canary pods.

Blue-Green Strategy

Blue-green rollout


apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: myapp
spec:
  replicas: 5
  strategy:
    blueGreen:
      activeService: myapp-active
      previewService: myapp-preview
      autoPromotionEnabled: false
      prePromotionAnalysis:
        templates:
          - templateName: blue-green-analysis
      postPromotionAnalysis:
        templates:
          - templateName: post-promotion-analysis
      scaleDownDelaySeconds: 300
      previewReplicaCount: 2
      antiAffinity:
        preferredDuringSchedulingIgnoredDuringExecution:
          weight: 100

Blue-green deploys the new version to a preview environment. Pre-promotion analysis verifies the new version before switching traffic. autoPromotionEnabled: false requires manual approval. scaleDownDelaySeconds keeps the old version running after promotion for quick rollback.

Analysis Templates

AnalysisTemplate for Prometheus metrics


apiVersion: argoproj.io/v1alpha1
kind: AnalysisTemplate
metadata:
  name: canary-analysis
spec:
  args:
    - name: service-name
  metrics:
    - name: success-rate
      interval: 2m
      count: 5
      successCondition: result[0] >= 0.99
      failureLimit: 3
      provider:
        prometheus:
          address: http://prometheus:9090
          query: |
            sum(rate(http_requests_total{service="{{args.service-name}}",status!~"5.."}[2m]))
            /
            sum(rate(http_requests_total{service="{{args.service-name}}"}[2m]))

    - name: latency
      interval: 2m
      count: 5
      successCondition: result[0] <= 500
      failureLimit: 3
      provider:
        prometheus:
          address: http://prometheus:9090
          query: |
            histogram_quantile(0.95,
              sum(rate(http_request_duration_seconds_bucket{service="{{args.service-name}}"}[2m])) by (le)
            ) * 1000

    - name: error-rate
      interval: 1m
      count: 3
      successCondition: result[0] <= 0.01
      failureLimit: 2
      provider:
        prometheus:
          address: http://prometheus:9090
          query: |
            sum(rate(http_requests_total{service="{{args.service-name}}",status=~"5.."}[1m]))
            /
            sum(rate(http_requests_total{service="{{args.service-name}}"}[1m]))

AnalysisTemplates define success/failure criteria using metrics from Prometheus. successCondition is a boolean expression — the analysis passes when it evaluates to true. failureLimit allows some failures before aborting.

Traffic Routers

Router	Use Case
NGINX Ingress	NGINX ingress controller with canary annotations
Istio	Service mesh with VirtualService traffic splitting
Linkerd	Service mesh with traffic split CRD
AWS ALB	Application Load Balancer with target group weights
Gateway API	Kubernetes Gateway API with HTTPRoute

Abort and Rollback

Manual abort and rollback


# Abort the current rollout
kubectl argo rollouts abort myapp -n production

# Retry the rollout
kubectl argo rollouts retry myapp -n production

# Promote to the next step
kubectl argo rollouts promote myapp -n production

# Promote to the final step
kubectl argo rollouts promote --full myapp -n production

# Check rollout status
kubectl argo rollouts status myapp -n production

# View rollout history
kubectl argo rollouts history myapp -n production

The CLI provides full control over the rollout lifecycle. abort stops the rollout and routes traffic back to stable. promote advances to the next step. promote --full skips all remaining steps and fully promotes the canary.

Use pause for manual gates

Add pause steps with no duration to create manual gates. The rollout pauses indefinitely until someone runs kubectl argo rollouts promote. This allows manual verification before traffic increases.

Mark this lesson complete to store local progress and unlock a cleaner resume path the next time you visit.

Stage 5 · Platform

GitOps & Progressive Delivery