In earlier versions of Zenoss if you received heatbeat failures and they were associated with a specific daemon I found that the rate of the incoming events overwhelmed the ability of the Zenoss system to process them. There were two approaches to addressing this issue 1. Get a more powerful/faster IO zenoss server. 2. Reduce the quantity of incoming events. In your case since the heartbeat errors are from zensyslog, I would suggest forwarding the syslog messages on your monitored devices to a syslog-ng install for "pre filtering" use this point to drop messages you don't care about so the Zenoss server will not have to do it. Note: like I said this was for earlier version, I have no experience with these fancy new versions but once I understood the Zenoss heartbeat mechanism, it actually turned out to be my friend.
↧