Icinga2 Comments Getting Recreated

I am trying to troubleshoot a high load average on machines in one of our zones that is causing some checks not to execute in a timely manner.

When investigating /var/log/icinga2/icinga2.log on either of the zone endpoints, I see thousands of entries similar to the follow (largely for the same service on the same host):

[2020-03-16 15:29:48 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!c51e4753-9756-4647-816c-1ac110e1c4cd' of type 'Comment'.
[2020-03-16 15:29:48 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!a3c6f92d-bc31-4fc4-90c4-3209dc79c08e' of type 'Comment'.
[2020-03-16 15:29:49 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!d7ec0747-1fa6-4cd7-993e-5e7b03cd9c25' of type 'Comment'.
[2020-03-16 15:29:49 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!2a9fde5a-82a2-40db-b443-ad81a83fca8b' of type 'Comment'.
[2020-03-16 15:29:50 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!04ae6654-1782-433c-bb42-46b3f9bf58ec' of type 'Comment'.
[2020-03-16 15:29:50 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!6bdb4eba-2d8f-4d7c-af33-32e39de1f84e' of type 'Comment'.
[2020-03-16 15:29:50 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!3ecc531d-f3ed-426b-8e10-508930c2a9df' of type 'Comment'.
[2020-03-16 15:29:51 -0500] information/ConfigObjectUtility: Created and activated object 'ar01.gbsntn01!BGP ESP STATUS 172.X.X.X!b1898131-82c7-4be5-b506-8049c5f5a493' of type 'Comment'.

This ballooned the icinga2 ido db, as icinga_comments, icinga_commenthistory, and icinga_acknowledgements all had 50k+ rows.

I cleaned up each of those tables, restarted icinga2 on the satellite agents. I still see the concerning log entries and high load average. Also, icinga_comments is getting filled up with 50K+ entries after the restart again, but acknowledgements and comment history seem to be fine.

What could be causing the comments to be created thousands of times?

Found in another forum post that I can clean up comments:
find $(cat active-stage)/conf.d/comments/ -maxdepth 1 -name "*.conf" -print0 | xargs -0 rm

This has helped some, letting Icinga run its course with some jobs to clean out the queue

1 Like

Hey there,
thanks for taking the time to post your solution, did that solve the problem or just put a plaster on the wound?

I’ll bump it over to the devs to look at it, maybe it’s something we should keep in mind!

1 Like