Hi Russell,
I think the first step here is to know which is the source of the high load of events which are flooding your agent. I have some questions to find it out:
- Do the flooding logs appear in every agent restart? This can mean that the FIM engine is monitoring too many directories (it performs a pre-scan when starting).
- Are the Windows audit monitoring noisy logs coming from the Exchange server? Maybe the event channel is generating a big amount of events related to Exchange, and the logcollector is monitoring all of them.
You can enable the <logall> option on the manager side to watch which events are coming from the flooded agents while the queue alerts appear. Raw events from the agent are stored at /var/ossec/logs/archives/archives.log when that setting is enabled.
On the other hand, you can tune the queue settings in the ossec.conf file (queue size and eps) and internal values (thresholds) at local_internal_options.conf. However, values which affect the performance of the queue are the ones located at ossec.conf.
If you find the queue is getting fully in exceptional times (when one of the components performs any scan), I suggest you increase the queue size to soften the burst of events. However, if the flooding is constant, I would increase the EPS of the queue as well as try to filter the noisy events.
I hope it helps.
Best regards,
Chema.
Chema Martinez | IT Engineer — Wazuh, Inc.