Deployment Monitor

The Deployment Monitor preset (wf-deployment-monitor) is integration-first: it routes on deployment outcome, then calls Slack (incoming webhook) when the deploy is healthy or GitHub (issue) when it failed — each path is guarded by hasIntegration:* so missing credentials never hard-fail the graph.

Flow

1. Deploy Webhook — ingress; payload is merged into workflow_run.context.
2. Normalize Deployment — reads context.deployment and sets severity to high when status is failed, failure, error, or rolled_back; otherwise low.
3. Deployment failed? — severity >= high.
   • TRUE → GitHub integration? → live GitHub: failure issue or stub.
   • FALSE → Slack integration? → Slack: deploy healthy or stub.

Sample context (“Run with sample inputs”)

The bundled sample uses a failed deploy so the GitHub branch exercises by default:

{
  "deployment": {
    "service": "payments-api",
    "status": "failed",
    "version": "v3.8.0",
    "environment": "production"
  }
}

To exercise the Slack path instead, run with deployment.status set to succeeded (or any non-failure status) and save a Slack incoming webhook plus (for the GitHub branch) a GitHub PAT and default repo on Dashboard → Integrations. See GitHub Integration (PAT) for the GitHub side.

Related

• Workflow catalog · Execution proofs