Skip to content
dagweavedagweave docs

Run on your cluster

dagweave never runs your workflows itself. Your own cluster’s Argo Workflows is always the runtime. dagweave submits runs to it and shows you the results: status, logs, metrics, and artifacts. This is the paid path. The free path is to build a workflow and take the YAML, which you apply yourself.

Two roles

Section titled “Two roles”

Connecting a cluster is a one-time job for one person. A cluster operator installs the connector once per cluster and turns on the features you want. After that, workflow authors run workflows from dagweave without touching Kubernetes.

More than one cluster

Section titled “More than one cluster”

You can connect more than one cluster. Install one connector in each. dagweave aggregates across all of them, and when you run a workflow or attach a trigger you choose which cluster and namespace it targets.

Teams run several clusters for real reasons:

  • Staging and production. Connect a test cluster and a production cluster, and promote the same workflow from one to the other.
  • Scale, region, or isolation. Split work across clusters by capacity, by region, or to keep tenants apart.

Two facts to keep straight. One connector serves one cluster; a single connector does not route across clusters. Within a cluster, Argo Workflows is often installed per namespace, so one cluster can host several Argo Workflows instances. You model each instance as an ArgoConnection, which the operator writes after installing the chart. The chart itself creates none, and the connector refuses to dispatch to a target no ArgoConnection matches.

Where to go next

Section titled “Where to go next”