Not alert on IIS app pool idle time but alert on shutdown

[yuting_wu]

We want to alert when an IIS app pool is not available in Dynatrace one agent . And we configure this in the process group level, availability tab -> if any process becomes unavailable. However, when the app pool goes to idle mode, it will still create a problem and send alerts. How do we configure this so that it will only alert on process shutdown but not on the idle time? Thank you!

[skrystosik]

It may be not possible that way, but you can change aproach. You can use synthetic monitor to call IIS application. If it will be in idle mode it will answer (but with longer time) if it will be shut down, dynatrace will rise an alert:

https://www.dynatrace.com/support/help/how-to-use-dynatrace/synthetic-monitoring/http-monitors/creat...

If your application is not available public you can create your own synthetic location:

https://www.dynatrace.com/support/help/how-to-use-dynatrace/synthetic-monitoring/private-synthetic/b...

Sebastian

[yuting_wu]

Can you elaborate more on how can we use HTTP monitor to detect idle time? It's not public

[dave_mauney · Accepted solution]

Calling with a synthetic will "wake up" the IIS app pool, resulting in it being up all the time. So why not just configure IIS to remove the 20 minute sleep setting? That would be free, and much less effort. Of course, you will waste CPU cycles with either of these approaches. Another option is to alert only when services are impacted (our default setting) since the lack of an IIS app pool is expected when traffic is not present.

[ankitmkgupta]

Hi @dave_mauney , do you think we can configure the alerts like

"Whenever App Pool recycle due to idle timeout or low traffic We should not get the alert.

FYI We are using windows docker and getting hundreds of alert on weekends.

Currently we are geting alert like below 

Process unavailable
Process IIS app pool XXXXXXX on host EC2XXZ-XXXX has been shut down

[dave_mauney]

I think your choices are to configure the process groups to never alert for process unavailable, to configure IIS to not stop the process when there is no traffic, to configure a maintenance window for weekends to avoid alerts, or to use the API to toggle the PG availability setting off for the weekend and on for the week days.  There may be other ways, but these are the ones that come to mind for me immediately.

I don't think there's a way for Dynatrace to know distinguish between normal IIS shut down of processes due to lack of traffic vs. the process being shutdown for other reasons.

[dambala_rosylan]

Hi Team,

Please let us know if solution is present now to differentiate the IIS processes that are really not available and the processes that were shutdown due to inactivity.

Regards,

Rosy