New events not processed or displayed

We’ve been running Sentry v20.6.0 (37a7530) for a while now and seemingly out of nowhere things started to act up.

  • Completely new issues are not appearing in Sentry
  • When a new event for an existing issue gets logged, the existing issue gets unresolved, but data about the new event is not added. Not even in the aggregated data. “Last seen” for example remains at “13 days ago” or so.
  • These missing events are present in the general organization stats, so they are getting picked up, but something goes wrong with the processing or displaying I guess.

Does anyone recognize this issue? How would we go about debugging this to find out what is going on?

2 Likes

Ok, I should have posted here sooner because an hour after I did so, the problem solved itself. The problem started 5 days ago and all of sudden an hour after posting here our projects in Sentry went from empty to full with events from the past 5 days :confused:

Looking at the issues in the internal project, here are the things I see that coincide with the time the problem started:

  • OperationalError sentry.tasks.options.sync_options - could not translate host name "postgres" to address: Temporary failure in name resolution
  • SnubaError /api/0/organizations/{organization_slug}/tags/ error - HTTPConnectionPool(host='snuba-api', port=1218): Read timed out. (read timeout=30)
  • A few times CancelledError /organizations/:orgId/issues/ - GET /organizations/mobilevikings/users/ 0, all for the same user
  • Exception sentry.tasks.check_auth_identity error - HTTP 500 (internal_failure): no description available

Does this mean anything to anyone?

Looks like a network hiccup or a temporarily overloaded system to me.