dagweave docs

What dagweave is

dagweave is the visual editor for Argo Workflows. You lay out a pipeline as a graph of nodes, wire the data and dependencies between them, and dagweave compiles that graph into standard Argo Workflows YAML. The output is the same shape you would write by hand, so it runs anywhere Argo Workflows runs.

It brings authoring Argo Workflows into the 21st century, so your team stops hand-writing YAML. dagweave does not host Argo Workflows and it is not an app-integration tool. You bring your own cluster, and dagweave authors and runs Argo Workflows on it.

Who it is for

Most people who need a workflow are not Kubernetes experts, and should not have to be. dagweave is for anyone who needs to build one without hand-writing YAML or learning the Argo Workflows schema first: a data scientist standing up a training pipeline, an engineer automating a job. You pick each step from a searchable list, connect them, and get a real manifest you own and can run.

Teams split into two roles:

• Workflow authors build workflows on the canvas. They never touch Kubernetes or install anything, and this is most people.
• Cluster operators install the connector once so authors can run those workflows on the team’s own Argo Workflows cluster.

Building and running

Building is the whole visual editor and it is not gated. Build any Argo Workflows workflow visually, import your existing manifests, edit nested workflows, wire data between steps, and copy or download the YAML. None of that needs a cluster or a connector.

These need a plan that includes them. The API answers 402 when yours does not, and the app names what is missing:

• Publish a workflow to a git repo
• Teams and multiple members
• Run workflows on your own cluster, with triggers (manual, cron, webhook, GitHub, AWS SNS) and testing a workflow against a real Postgres row or S3 object
• The executions view with failed-run root cause and re-run
• Run metrics
• Browse and install shared workflows from the template catalog
• Fine-grained access control
• An audit log
• Publish your own catalog listing

Your plan also sets how many workflows you can keep, and whether you get the full node set or the core one.

Dispatching a run needs the connector installed in your own cluster on top of the plan. Without it you build, import and publish YAML, and nothing runs.

Where to go next

• Getting started walks you from an empty canvas to real Argo Workflows YAML you own.
• Core concepts explains the canvas, how it compiles, how you wire data between steps, and how import round-trips your existing YAML.
• Installing the connector covers the in-cluster operator that lets dagweave submit workflows and collect results.
• Triggers covers running a workflow manually, on a schedule, or from an event.
• Executions covers watching runs, reading logs, and re-running.
• Test with real data covers testing a workflow against a real Postgres row or S3 object from a connected source.
• Teams and access control covers teams, roles, fine-grained access, MFA, and the audit log.