Problem with Zabbix performance - History syncer processes 100%

[gcaselli]

Hi all,

I'm in the process of setting up a new Zabbix environment. This Zabbix instance is containerized within Azure Kubernetes Services (AKS), currently monitoring 249 hosts and 32,928 items, with 6 proxies in operation.

However, I'm encountering an issue where the process history syncer consistently hits 100%, causing a slowdown in monitoring. Interestingly, restarting the container temporarily resolves the issue, but it resurfaces after a while.

Here are some additional details:

• Zabbix Backend and Frontend are containerized in AKS, running version 6.4.10.
• The database is also containerized in AKS, using MariaDB 11.0.3 managed via Kubernetes Operators for MariaDB. Kubernetes Operators for MariaDB - MariaDB Knowledge Base
• Zabbix Proxies are installed on-premise K3S clusters, also running version 6.4.10.

System Information:

Number of hosts	249
Number of items	32928
Number of triggers	19228
Required server perf	371.29

Zabbix Backend non default values:

- ZBX_HISTORYSTORAGEDATEINDEX: "1"
- ZBX_STARTREPORTWRITERS: "3"
- ZBX_DEBUGLEVEL: "3"
- ZBX_JAVAGATEWAY_ENABLE: "false"
- ZBX_ENABLE_SNMP_TRAPS: "true"
- ZBX_CACHESIZE: "512M"
- ZBX_VALUECACHESIZE: "512M"
- ZBX_HISTORYCACHESIZE: "128M"
- ZBX_HISTORYINDEXCACHESIZE: "64M"
- ZBX_STARTDBSYNCERS: "7"
- ZBX_TIMEOUT: "4"
- ZBX_LOGSLOWQUERIES: "3000"
- DEBUG_MODE: "false"


I'm seeking advice on how to improve performance in this scenario.

Thanks a lot!

Best Regards,
Gianni.

[gcaselli]

in AKS, Zabbix is in a kubernetes pool with 1 node, and the Database is in a pool with 2 nodes, the nodes are of size Standard_D4ds_v5 in Azure.