Inspect a run

Start with the Run snapshot:

helmr run get RUN_ID --project agents --env development
helmr run wait RUN_ID --project agents --env development --timeout 10m

The snapshot includes status, entrypoint, Deployment, Workspace, current attempt, cause, metadata, tags, timestamps, and terminal output or failure.

Use logs for process output and structured application logging:

helmr run logs RUN_ID --project agents --env development
helmr run logs RUN_ID --project agents --env development --follow

Use events for lifecycle decisions such as waits, retries, cancellation, and finalization:

helmr run events RUN_ID --project agents --env development

The SDK exposes cursor pages and filters:

const run = await client.runs.retrieve("RUN_ID")
const logs = await client.runs.logs(run.id, {
  level: ["warn", "error"],
})
const events = await client.runs.events(run.id, {
  severity: "error",
})

Actor application output is not a Run log. Read it from the associated Session when run.sessionId is present.