Zenoss' zenperfsnmp daemon generates a lot of events. In most cases it is the leading source of events by a significant margin.
Depending on the monitoring templates in place and the number of devices being monitored by Zenoss, zenperfsnmp may be raising thousands of events during each cycle. Before being processed by the event engine these events are held in a queue, the length of which is determined by the config parameter maxqueuelen.
If the queue of events exceeds maxqueuelen, new events are dropped indiscriminately. This is obviously undesirable, even if it happens only occasionally. But when your zenperfsnmp event queue looks like this...
...you're likely to be consistently dropping events.Continue Reading Comments