Argo Rollouts is a Kubernetes controller and set of custom resource definitions that extends the native deployment capabilities of Kubernetes with advanced progressive delivery strategies. Written in Go, the project provides blue-green deployments, canary releases, canary analysis, experimentation, and progressive delivery features that address significant limitations in Kubernetes' standard RollingUpdate strategy. The standard rolling update approach lacks fine-grained control over rollout speed, provides no traffic shaping capabilities, cannot query external metrics for validation, and cannot automatically abort and rollback failed updates, making it unsuitable for large-scale production environments where blast radius control is critical.
The core functionality of Argo Rollouts centers on traffic management and automated decision-making during deployments. The controller integrates with multiple ingress controllers including NGINX, ALB, Apache APISIX, Ambassador, Traefik, Contour, and Gateway API, as well as service meshes like Istio, Linkerd, and SMI. These integrations enable weighted traffic shifting, mirroring, and header manipulation capabilities. The project supports metric provider integration with Prometheus, Wavefront, Kayenta, Web endpoints, Kubernetes Jobs, Datadog, New Relic, and InfluxDB, allowing deployments to be automatically promoted or rolled back based on key performance indicators and business metrics.
The repository shares contributors with related projects including argoproj/argo-cd, argoproj/argo-workflows, and microsoft/vscode, reflecting its position within a broader ecosystem of deployment and workflow automation tools.
Argo Rollouts provides manual judgment capabilities alongside automated decision-making, allowing operators to pause deployments for manual verification before proceeding. The fine-grained weighted traffic shifting enables gradual rollout of new versions with precise control over the percentage of traffic directed to each version. Automated rollback functionality triggers when metrics fall below defined thresholds, protecting production environments from degraded deployments. The project is classified across multiple domains including blue-green deployments, canary releases, traffic splitting, traffic management, traffic shifting, rollout analysis, deployment automation, observability, and continuous delivery, reflecting its comprehensive approach to progressive delivery.
The documentation is maintained at argo-rollouts.readthedocs.io and includes getting started guides and complete feature documentation. The community maintains active engagement through GitHub Discussions, Slack channels, weekly contributors office hours on Thursdays, and monthly user community meetings. The project has been adopted by numerous organizations documented in an official user list, and the community has produced extensive content including blog posts, video presentations, and case studies covering topics from basic canary deployments to multi-service progressive delivery and stateful service deployments.
