Prom metric for "failed" doesn't account for recently resumed tasks

When resuming a paused task, the prom metric will report "failed" until next scheduled run, and not take into account that the task was recently unpaused.

Locate the Prometheus metric logic and the code path that resumes a paused task, then trace how its failed status is determined before the next scheduled run. Confirm the expected metric state immediately after unpausing and add or update coverage if an existing test area is found; done means a recently resumed task is no longer reported as failed.

Written by the indexing model from the issue text.