Zabbix SQL server 100% useage and webinterface crashes

[MikeJensen]

I have a problem with our Zabbix 7.0.9 environment.
We have Server A which is our Zabbix Frontend Server, and Server B which is our Zabbix SQL server with TimescaleDB.

Sometimes the Zabbix server throws extremely long queries after the SQL server, and then it crashes.
Eg. today from 20250428:160005.386 to 20250428:160020.020 (1 query every second) i get 5 warnings in the Zabbix server log: each is "slow query: 3-5 seconds,"
Example: slow query: 3.540515 sec, "update trends_uint set num=.......
Each of these queries are between 60k and 80k characters long, which makes our SQL server use 100% CPU at all times.

We have as of now 677 hosts.

This wasn't a problem before we migrated from 6.0 to 7.0.

Does anyone have a clue what could be our issue?

[mrnobody]

According manual page (and tests i have done), unique change necessary is log_bin_trust_function_creators = 1 BEFORE using SQL import scripts, and after that, just rollback = 0. Did you do that?

But with TimescaleDB, i don't know, probably something more in necessary.

[cyber]

According manual page (and tests i have done), unique change necessary is log_bin_trust_function_creators = 1 BEFORE using SQL import scripts, and after that, just rollback = 0. Did you do that?

That is only required for mysql/mariadb, not Postgres/timescale

MikeJensen Is that a all-in -one instance? I guess your host is just out of steam... For starters give it more memory. What kind of disks you are using? Faster is better (ssd)...

[MikeJensen]

cyber
I have a 3 node setup, A Zabbix-Server on one node, Zabbix SQL on a 2nd node and a Zabbix-Proxy on a 3rd node, with approx equally host distribution on Server/Proxy.
All servers are Virtual Machines hosted on a Azure HCI environment. The Server and SQL both have 32gb memory, the proxy have 16gb.
The Server and proxy are on Ubuntu 22.04.1 and the SQL is on Redhat 11.5.0 (We want to move the SQL server to Ubuntu 22.04.1, but we have trouble with this at the moment)

I've been sick for the last week, but i can see in the load average for the SQL server last 10 days, that it have had 2 days with high load. Both time starting at around 10-11 in the morning.

[tim.mooney]

I'm not 100% certain your issue is ZBX-25651, but it sounds suspiciously like it.

See the list of fixed issues in the release notes for 7.0.11: https://www.zabbix.com/rn/rn7.0.11

The moral of the story here is that when you're having an issue, always make sure you're running the latest version in your series. 7.0.11 isn't even the latest now.

[MikeJensen]

tim.mooney i have now upgraded from 7.0.9 to 7.0.12, i will se how it behaves in the comming days.