Zabbix configuration syncer processes more than 85% busy

[mashuma]

Hello,

I've got following issue that I can't wrap my head around.. occasionally we are getting this alarm.

What is configuration syncer ?
How would one go about troubleshooting this ?

I did not manage to find info on this in the docs

Thanks a lot in the advance!

[mashuma]

To illustrate the issue, here are the screenshots of the "Zabbix busy configuration syncer processes, in %".

[warp10]

mashuma Hi , I think you look this post https://www.zabbix.com/forum/zabbix-...e-than-75-busy and this one https://ma.ttias.be/startdbsyncers-zabbix/

[mashuma]

Thank you!
However the mentioned posts are discussing the DB syncers, not the configuration syncers. The DB syncer processes are not busy at all, looking at the zabbix performance graphs.
Could this be DB issue ?

As for the zabbix server configuration, I've got these values

StartPollers=400
StartPollersUnreachable=30
StartTrappers=15
StartPingers=30
StartDiscoverers=30
StartHTTPPollers=15
StartJavaPollers=20
StartSNMPTrapper=1
CacheSize=300M
StartDBSyncers=12
HistoryCacheSize=80M
HistoryIndexCacheSize=16M
TrendCacheSize=100M
ValueCacheSize=96M
Timeout=10
StartProxyPollers=15

Please see attached screenshot of
- Zabbix internal process busy %
- Zabbix server performance
- Zabbix data gathering processes busy %
- Zabbix cache usage %

[mashuma]

+----------------------------------------------------+---------+--------------------------+
|                     Parameter                      |  Value  |         Details          |
+----------------------------------------------------+---------+--------------------------+
| Zabbix server is running                           | Yes     | localhost:10051          |
| Number of hosts (enabled/disabled/templates)       | 497     | 315 / 55 / 127           |
| Number of items (enabled/disabled/not supported)   | 111478  | 61570 / 38577 / 11331    |
| Number of triggers (enabled/disabled [problem/ok]) | 16249   | 12043 / 4206 [6 / 12037] |
| Required server performance, new values per second | 1026.61 |                          |
+----------------------------------------------------+---------+--------------------------+

[warp10]

mashuma Could you tell us about your zabbix architecture and hardware ? are you disable your item not supported ?

[mashuma]

Zabbix server version 4.2.4 with MariaDB 10.2.25

Server and DB server are different VMs on different hypervisors.
Not sure how detailed I need to get with this. If it is suspected it is DB issue, then I can provide more stats.

[warp10]

mashuma Hi Check your "interval" Time of all your Item , for your config your syncer is about 24% / 25% perhaps less.
push your StartDBSyncers=12 to 20 wait and see the max is 100, but for your information too many have a négative effect on performance,
Check the trigger too if you have not a positive fault with your message syncer process.

[ingus.vilnis]

Poor database performance is the top suspect in such cases. Pushing the DB syncers to 20 will make it even worse. Default 4 DB syncers are more than enough for your instance @ 1000 nvps. Besides, DB syncers have nothing to do with the configuration syncer which is the topic of this thread.

Check if you have LogSlowQueries=3000 enabled in your zabbix_server.conf and check for slow queries in your Zabbix server log. Check under which PID is the configuration syncer running on your server and grep for this number in the log.

[mashuma]

Thank you, ingus.vilnis . Will enable slow logs and focus on DB performance metrics. Will report back, when I find more clues.