monitoring for failed polling engine

FormerMember

Does anyone know of a tool or devised an automated way to detect when the polling service stops polling? We have multiple pollers and this happens very infrequently and randomly amongst pollers. We go weeks/months between incidents. But, when it does we don't realize it until a NOC person or technician happens to select a node assigned to the failed poller and receives "no data for selected period" message instead of a graph. Sometimes it has been several days before someone realizes a poller has died. Our only resolution is to kill the NetPerMonService and restart it and then it runs weeks or months before another incident. Frustrating! We monitor the NetPerfMonService but that isn't good enough because the service continues to run, the poller just stops sending polls. Does anyone know if the NetPerfMonService is really still running or is it hung? If it is still running it would be helpful if the service could detect polling has stopped and log an Orion event or syslog. We can then trigger alerts/pages or automated action (kill & restart service) to restore polling in a more timely manner. Wishful thinking huh?

Janis1

SolarWinds sells a Hot Standby Management Pack. On the website it states the following: "The Orion Network Performance Monitor Hot Standby Engine will consistently check the Primary systems’ ability to function and record availability and performance data. In the event of a failure, the Hot Standby Engine will automatically assume network-monitoring responsibility. When a failed Polling Engine returns to an operational state, the Hot Standby Engine will automatically return all monitoring tasks to the original Polling Engine". More information is available at www.solarwinds.net/.../overviewHSMP.htm. This looks like it is exactly what you are looking for.

Janis