topic Max process count exceeded unreasonably in Splunk AppDynamics

Max process count exceeded unreasonably

Trent_Kerin — Wed, 04 Aug 2021 02:33:12 GMT

Hi I've got some machine agent installations where I'm getting messages like this:

[#|2021-08-02T14:38:39.254+1000|WARNING|glassfish 4.1|com.appdynamics.SIM|_ThreadID=80;_ThreadName=http-listener-2(7);_TimeMillis=1627879119254;_LevelValue=900;|#SIM000121 The maximum number of monitored processes per machine allowed has been reached for machine 33. The limit sim.processes.count.maxPerMachine is set to 1000 processes. This limit will be reset after the next process purging or when some processes are deleted by the user. Could not create 9 processes for machine 33|#]

I then go look at machine 33, and find that it's got numerous duplicates of the same processes, varying only in start/end time.

It seems like if I increase the maxPerMachine limit, we'll just delay running into the limit again because it's constantly using up the count for the same processes over and over.

This seems like a bug. Is there some workaround?

Re: Max process count exceeded unreasonably

Claudia_Landiva — Wed, 04 Aug 2021 21:25:08 GMT

Hello, @Anonymous

You're right: increasing the limit would just delay running into the limit without addressing the root cause. Since the agent team would need to review your logs to get to the bottom of the issue, it's best that you open a Support ticket.

Thank you,

Claudia Landivar

AppDynamics Community Manager & Editor

Re: Max process count exceeded unreasonably

iamryan — Mon, 09 Aug 2021 21:32:51 GMT

Hi, @Trent.Kerin

If you happen to create a support ticket and get info from Support, please share those learnings back as a reply to this post. Knowledge sharing is key to the growth and success of the community.

Re: Max process count exceeded unreasonably

Trent_Kerin — Tue, 10 Aug 2021 01:34:18 GMT

I got a reply back from support:

The duplicate processes behaviour is a known product bug that's "under PM evaluation".

The workaround for now is to exclude processes where possible to try and avoid running back up to the limit. I'll need to check out what works for my servers.