fix(slack): Handle archived channels gracefully instead of logging errors#256
Conversation
…rors Catch the Slack API 'is_archived' error in all channel-mutating methods (join, post, pin, bookmark, invite, rename, set topic) and log a warning instead of an error. Also add an upfront archived-channel check in the backfill handler to skip setup entirely and avoid sending a misleading DM about inviting the bot. Include is_archived in get_channel_info return value and reuse the existing get_channel_info call in backfill setup to avoid a duplicate API request. Fixes FIRETOWER-BACKEND-4D Co-Authored-By: Claude <noreply@anthropic.com> Agent transcript: https://claudescope.sentry.dev/share/iSJ8GLrggJu2yvWAMY0Qg8VvHJ9AgfdC3lL8L3cly6A
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
![sentry[bot]](https://avatars.githubusercontent.com/in/12637?s=60&v=4){kind=small}
…etry When post_message or add_bookmark hit a not_in_channel error and the subsequent join_channel call fails, the failure was silently swallowed. Add explicit error logging for these cases. Co-Authored-By: Claude <noreply@anthropic.com> Agent transcript: https://claudescope.sentry.dev/share/WUxpRlQ-Dvgz-R6V962vnuF0JTfCMODq8t1uPCJ65HY
When the initial get_channel_info call returns None (transient API error) but join_channel succeeds, the rename step was silently skipped because it short-circuited on the None channel_info. Re-fetch channel info after a successful join when the initial lookup failed so the rename proceeds correctly. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Agent transcript: https://claudescope.sentry.dev/share/RgcHeFRWgaj2R6OxQjKoWYR9gJhMlO3Hj9fat4C-AmA
![sentry-warden[bot]](https://avatars.githubusercontent.com/in/2755838?s=60&v=4){kind=small}
The side_effect list had only 2 entries: one consumed by handle_backfill_submission and one by _setup_channel_for_incident's initial get_channel_info call. Because the second call returned a valid dict, the `if not channel_info` retry branch was never reached. Add a third side_effect entry so setup's initial call also returns None, forcing the retry path after join. Update expected call_count from 2 to 3 accordingly. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Agent transcript: https://claudescope.sentry.dev/share/SZpW242OEy3gh6Vzune4dgeQjtHQ3ESeFLcWvlszNfo
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 950413f. Configure here.
| channel_info = _slack_service.get_channel_info(channel_id) | ||
|
|
||
| expected_name = build_channel_name(incident) | ||
| channel_info = _slack_service.get_channel_info(channel_id) |
There was a problem hiding this comment.
Ignores archived flag from submission
Medium Severity
During backfill submission, get_channel_info runs before _setup_channel_for_incident, but setup fetches channel info again and only uses that second result for the archived check. If the first call returned is_archived and the second returns None, setup skips the early exit, join_channel fails, and the user still gets the invite-the-bot DM the change was meant to avoid.
Additional Locations (1)
Reviewed by Cursor Bugbot for commit 950413f. Configure here.
| logger.warning( | ||
| "Skipping setup for archived channel %s on incident %s", | ||
| channel_id, | ||
| incident.id, | ||
| ) |
There was a problem hiding this comment.
Incident created silently with no user notification when channel is archived during backfill submission
When handle_backfill_submission is called for an archived channel, the incident is persisted to the DB but _setup_channel_for_incident returns early at line 57 without sending any DM, leaving the user with no confirmation the incident was created or how to find it.
Evidence
handle_backfill_submissioncallsserializer.save()to create the incident unconditionally, then calls_setup_channel_for_incident.- The new early-return block (lines 53–57) logs a warning and returns with no call to
client.chat_postMessage. - The only user-visible completion message is sent at the bottom of
_setup_channel_for_incidentafter all setup steps — it is never reached for archived channels. - The test
test_skips_setup_for_archived_channelexplicitly assertsclient.chat_postMessage.assert_not_called(), confirming this silent behavior is the current design. - Before this change,
join_channelreturnedFalsefor archived channels and a DM was sent telling the user to invite the bot — imperfect but at least actionable feedback was delivered.
Identified by Warden find-bugs · EZM-YQ3
2 participants
Catch the Slack API 'is_archived' error in all channel-mutating methods (join, post, pin, bookmark, invite, rename, set topic) and log a warning instead of an error. Also add an upfront archived-channel check in the backfill handler to skip setup entirely and avoid sending a misleading DM about inviting the bot.
Fixes FIRETOWER-BACKEND-4D
Co-Authored-By: Claude noreply@anthropic.com
Agent transcript: https://claudescope.sentry.dev/share/iSJ8GLrggJu2yvWAMY0Qg8VvHJ9AgfdC3lL8L3cly6A