guardian.py has a roundabout way of finding the owner of a given queue: it looks for any active connections (normally only one) and sets the user of that connection as the queue's owner. If it doesn't find one (client hung up right away or can't stay connected for long), the queue is orphaned, and no one is emailed about it if it gets too big and/or is deleted (at least until bug 832415 is fixed). However, this was conceived before we settled on a restrictive naming convention. Now you can easily tell the owner of a queue (or an exchange, see bug 1079523) from its name, e.g., queue/<username>/... We might as well just use that method in guardian.py.
This bit me again today due to my terrible memory. Going to fix this.
Assignee: nobody → mcote
Status: NEW → ASSIGNED
Fixed and added test for abnormal queue names. Did a bunch of cleanup at the same time. https://github.com/mozilla/pulseguardian/commit/3b62a48e577f82d1f01625d977cc157a57b9f48e
Status: ASSIGNED → RESOLVED
Last Resolved: 4 years ago
User Story: (updated)
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.