HTTPs Monitor Timeout

mcbridea

Hi All,

I have an application that is defined only by about 30 HTTPS monitors. Most of the monitors work but when one of the first few montors times out the lowest few monitors error out with the message

"Connection timeout. Previous monitor(s) used whole application job time. No time to run this one."

I have extended the application polling timeout from 5 sec to 20 sec but I am not sure this is the setting that will correct this issue. Is this the correct place to fix this issue or is there another setting for the total allowable monitor time?

Thanks - Andy

Affinion Group

aLTeReGo

Andy, this is indicative of the fact that there are too many component monitors within the application template to complete polling within the polling timeout window specified within the application setting. Component monitors that are successful complete in a more timely fashion, while those that fail in error take longer to complete as there are multiple retries made for each failed component. The Polling Timeout value is defined for the entire application/template, so the more components within the monitored application the longer the Polling Timeout and Polling Frequency values should be. There's no hard/fast rule for the number of components in an application for how long the timeout/frequency values should be since each component monitor type has its own unique overhead and retry methods. Network latency between the poller and the monitored host also plays a large part for how long a polling cycle takes to complete. I suggest increasing the Polling Timeout values each to 500 seconds. This should resolve the issue you are experiencing.