Argo Workflows
Argo Workflows is a Kubernetes-native workflow engine that runs each step of a pipeline in its own container, letting users define multi-step CI/CD, data-processing, or ML jobs as DAGs or sequential steps using Kubernetes custom resources. It belongs to the same CNCF Argo project family as Argo CD.
Baseline topologies
| Topology | Minimum nodes | Roles |
|---|---|---|
| single-replica | 1 | argo-controller, argo-server |
| ha | 2 | argo-controller, argo-server |
What drives the cost
- Peak concurrent workflow executions
- Peak number of workflows executing simultaneously. Each workflow can spawn many Pods; this drives Workflow Controller memory and Kubernetes API call rate.
- Peak concurrent Pods per workflow
- Maximum number of Pods a single workflow launches in parallel (DAG fan-out width). Combined with concurrent executions, determines peak Pod count the Kubernetes cluster must accommodate.
Where it runs
kubernetes
Size Argo Workflows and get a cost estimate