Spinnaker Orca task stuck / incomplete in CI

Orca is Spinnaker's orchestration engine that drives each task in a stage. When a task stays RUNNING forever or fails, Orca could not advance it, usually because a service it calls (Clouddriver, Front50, Igor) did not respond as expected.

What this error means

A stage sits in RUNNING indefinitely, or Orca logs show a task that will not complete. The execution never progresses past a particular task.

Spinnaker

orca: task 'monitorDeploy' in stage 'Deploy' did not complete;
retrofit error reading Clouddriver#getTask(...): read timed out

Common causes

A downstream service is slow or unreachable

Orca polls Clouddriver/Front50/Igor; if one is down or throttled, the task retries without completing.

The task genuinely errored but was retried into a stall

A transient error kept the task in a retry loop rather than surfacing a clean failure.

How to fix it

Check the service Orca is polling

Read Orca logs to see which downstream call is timing out.
Confirm that service (Clouddriver, Front50, Igor) is healthy and reachable.
Restore the service, then cancel and re-run the pipeline from CI.

Cancel and restart the execution

Once the downstream service is healthy, cancel the stalled execution and re-trigger so Orca starts a clean task.

How to prevent it

Monitor health of Clouddriver, Front50, and Igor.
Set sane timeouts so stalled tasks fail instead of hanging.
Scale Orca and its Redis backing store for pipeline volume.