argoproj/argo-workflows

Argo Workflows is an open-source tool that lets you run automated, multi-step jobs on Kubernetes (a system for managing software containers, think of it as an orchestrator that keeps your apps running across many computers). Instead of manually triggering one task after another, you define a workflow, a sequence of steps or a dependency map, and Argo handles the scheduling, execution, and monitoring automatically. Each step in a workflow runs inside its own container (a lightweight, isolated environment), which makes the system cloud-agnostic: it works on any Kubernetes cluster regardless of which cloud provider you use. You can model workflows as simple sequences or as a DAG (directed acyclic graph), describing which tasks must finish before others can start, useful when some steps can run in parallel. It is used for machine learning pipelines, data and batch processing, CI/CD automation (building and deploying code), and infrastructure automation. The project is a graduated member of the Cloud Native Computing Foundation (CNCF), meaning it has met rigorous maturity and adoption standards. Key features include a visual UI to monitor running workflows, support for storing files (artifacts) from cloud storage services, scheduled workflows via cron (time-based triggers), retry and timeout controls, REST and gRPC APIs, and single sign-on via OAuth2/OIDC. Client libraries are available in Java, Go, Python (via Hera), and TypeScript (via Juno). Written in Go, with over 200 organizations officially using it.