Zabbix Server Stucks every hour for few minutes

[rijesh.tp]

Hi Everyone,
Im facing a weird issue with my zabbix that it keeps stuck every hour for few minutes,
During that it sends me false alerts , saying client machines are not reachable.
After 5 to 10 minutes it all back to normal.
Please see the attached graph for zabbix performance during this time.
Thanks in advance for the help.

[dimir]

Is it your Housekeeper deleting old data every hour?

[kloczek]

Is it your Housekeeper deleting old data every hour?

Only if housekeeper is enabled and it starts every hour.
Second possible cause is that every hour are calculated rend points and this is causing some additional punch of write operations to the DB backednd and that additional traffic clogs DB backends.
Monitoring of the DB backend activity should show what is causing that issue. If delete queries -> housekeeper. If moire inserts -> trends points calculations.
In both cases improve DB backend will solve cause.
If issue is caused by housekeeper moving to partitioned history/trends tables will improve DB bakckend.

[mazofeifa]

Has anyone solved this issue? I have the same issue. Every hour between XX:10 and XX:15 the bandwidth graph for my network stops working. I don't see any log on the Zabbix service or MySQL service.

[cgillett]

Hi Rijesh, we are experiencing a very similar issue in Zabbix. Did you ever find a solution to this issue?

[tim.mooney]

Hi Rijesh, we are experiencing a very similar issue in Zabbix. Did you ever find a solution to this issue?

As Dimir suggested, the first thing to look at is whether the outages correspond to times when the HouseKeeper is active.

You didn't tell us anything about your environment, but assuming you haven't changed the default 'DebugLevel' setting on your server, you're probably getting log messages in your zabbix_server.log when the housekeeper runs. Check whether those overlap with the outages. If they do, then some database tuning is probably in order.