Verify a run
After the control and dispatcher services are ready and at least one worker is active, verify the environment with a small run. A default quickstart stack must enable NAT and worker capacity before this step.
Log in to the control URL:
helmr login "$CONTROL_URL"
Deploy the task project from a local checkout:
helmr deploy .
Start a small session and watch the run attempt reach a terminal state:
helmr session start TASK_ID --wait --follow
helmr run list
helmr run logs RUN_ID
A complete smoke test proves that:
- GitHub OAuth is configured.
- The control plane can read project metadata.
- A worker is active and can lease work.
- The worker can reach S3, ECR, AWS APIs, the control plane, and any external services used by the task.
- Firecracker, BuildKit, CNI, and guest artifacts are present on the worker.
If a run stays queued, check worker capacity first. If a task cannot reach an external repository or API, check the task secret and worker egress path.