← back

📷 "Little abacus in the server room" by Gael Varoquaux is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/.

GitOps Automation 2026: ArgoCD Image Updater and Rollouts in Tandem

08 August 2026 · 4 min · Martin Jochum #GitOps#ArgoCD#Argo Rollouts#Image Updater#Kubernetes#DevOps#Progressive Delivery

GitOps with ArgoCD has evolved in recent years from a pure synchronization tool into a comprehensive automation platform. While the comparison between ArgoCD and Flux in June 2026 shows that ArgoCD is the more popular choice with around 23,100 GitHub stars and version 3.4.3 (Tech Insider, 2026), the real added value for many teams lies in two specialized extensions: ArgoCD Image Updater for automatic container image updates and Argo Rollouts for progressive delivery. This article shows how both tools work together to enable a fully automated GitOps pipeline.

ArgoCD Image Updater: From CI Build to Git Commit

The ArgoCD Image Updater (current version 1.2.1) is a separate controller that monitors container registries and automatically updates the image tags in the Git repository when new images are available. This closes the gap between the CI pipeline and GitOps: a built container image lands in the registry, the Image Updater detects the new version and commits the updated tag back to the Git repository, and ArgoCD synchronizes the change into the cluster.

Setup is done via Helm chart or kubectl and takes just a few minutes:

helm repo add argo https://argoproj.github.io/argo-helm
helm install argocd-image-updater argo/argocd-image-updater -n argocd

The Image Updater is activated per ArgoCD application via annotations:

annotations:
  argocd-image-updater.argoproj.io/image-list: myapp=ghcr.io/myorg/myapp
  argocd-image-updater.argoproj.io/myapp.update-strategy: semver
  argocd-image-updater.argoproj.io/write-back-method: git
  argocd-image-updater.argoproj.io/git-branch: main

Update strategies at a glance: The Image Updater supports four strategies. semver selects the highest compatible semantic version tag, newest-build uses the most recent build date, alphabetical sorts alphabetically, and digest tracks a specific tag and updates on digest changes—for example, for latest tags. For Helm-based deployments, the values paths to be updated (helm.image-name, helm.image-tag) can be specified precisely; with Kustomize, the kustomization.yaml image object is updated (OneUptime, January 2026).

The write-back-method: git ensures that every change lands in the Git repository as a clean commit with a complete audit trail—a central GitOps principle. For teams that want additional quality assurance, a branch strategy is recommended: the Image Updater commits to a separate branch, and an automated pull request workflow enables reviews and automated tests before merging.

Argo Rollouts: Progressive Delivery with Automated Rollback Decisions

While the Image Updater ensures current images, Argo Rollouts handles safe delivery. Argo Rollouts (approx. 3,500 GitHub stars) extends ArgoCD with advanced deployment strategies that go far beyond the Kubernetes standard rolling update (OneUptime, February 2026).

The core is replacing the native Kubernetes Deployment resource type with a Rollout featuring a detailed canary or blue-green strategy:

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: myapp
spec:
  replicas: 5
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
      - name: myapp
        image: myapp:1.0.0
  strategy:
    canary:
      steps:
      - setWeight: 10
      - pause: { duration: 5m }
      - setWeight: 30
      - pause: { duration: 5m }
      - setWeight: 60
      - pause: { duration: 5m }
      - setWeight: 100

However, Argo Rollouts truly shines through Analysis Templates. These define metric queries—typically against Prometheus—and let the system automatically decide between promote or rollback (OneUptime, February 2026):

apiVersion: argoproj.io/v1alpha1
kind: AnalysisTemplate
metadata:
  name: success-rate-check
spec:
  metrics:
  - name: success-rate
    interval: 2m
    successCondition: result[0] >= 0.95
    failureLimit: 3
    provider:
      prometheus:
        address: http://prometheus.monitoring.svc:9090
        query: |
          sum(rate(http_requests_total{status=~"2.*"}[5m]))
          / sum(rate(http_requests_total[5m]))

If the success rate falls below 95% or the p99 latency exceeds a defined threshold, Argo Rollouts automatically aborts the canary and scales the new version back—without human intervention.

Practical Integration: From Push to Production Rollout

The combination of both tools results in a fully automated pipeline without manual steps:

  1. CI/CD pipeline builds a new image and pushes it to the registry (e.g., ghcr.io/myorg/myapp:v1.2.3)
  2. ArgoCD Image Updater detects the new tag, checks the semver rules, and commits the change to the Git repository
  3. ArgoCD synchronizes the updated rollout into the cluster
  4. Argo Rollouts starts the canary strategy—10% traffic, metric monitoring, gradual increase
  5. On metric violation: automatic rollback—otherwise full promote

This workflow corresponds to the “Intermediate” setup recommended in the DevOps community for teams with 10–50 services: Renovate for dependency updates, ArgoCD Image Updater for image promotion, and Argo Rollouts for automated canary deployments (DEV Community, 2026). A PodDisruptionBudget and correctly configured readiness probes remain indispensable as a safety net.

Conclusion

ArgoCD Image Updater and Argo Rollouts make ArgoCD 2026 more than just a pure GitOps sync tool. The Image Updater automates the tedious manual tag update process, while Rollouts enables progressive delivery with automated, metric-based rollback decisions. Together, they form a pipeline that runs completely automated from CI push to production rollout—while maintaining the central GitOps principle: Git remains the single source of truth. For teams already using ArgoCD, both extensions are a logical next step toward a fully automated deployment platform.

Sources

🌐 Machine-translated from the German original, editorially reviewed. 🤖 Written with AI assistance.