-
Checks random unknow status timedout
Hello Team,
It happens that randomly cheks go into unknow state. Regardless of the type of check (WMI, plugin, snmp …) I see in the server logs that random checks go into timed out
Are there perhaps parameters in the padnora_server.conf file to handle the concurrent number of checks that the pandora server can handle ?
Thank you very much for the support
-
2 Replies
-
Hello,
Thank you for your query.
What you describe are modules in unknown state due to timeouts in random checks (WMI, SNMP, plugins, etc.) this is usually related to an overload of concurrent checks that Pandora FMS server can’t manage efficiently.
In the pandora_server.conf configuration file, you can adjust several parameters to improve the server performance in these situations:
Key parameters you can check:
network_threads: Number of threads for remote checks (ICMP, TCP, etc.).
If you have many network modules, leave it at 8.
If not, 5 is enough.
plugin_threads: Threads for remote plugin execution.
If you use many plugins, you can raise it to 8.
If not, 5 is fine.
wmi_threads and snmp_threads:
For both, 5 is usually a suitable value.
multiprocess: This parameter allows multiple instances of the network server to be launched in parallel, which can improve performance.
You can set this multiprocess parameter to 1 in order to activate this improvement.
In case you don’t see this parameter just add it yourself at any point of the pandora_server.conf file
network_threads 5
plugin_threads 5
wmi_threads 5
snmp_threads 5
multiprocess 1
Remember to restart the service after applying the changes:
systemctl restart pandora_server
Best regards
Diego
-
