cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

This product reached the end of support date on March 31, 2021.

Multiple .net processes with same name on the same host

joe_wells
Participant

Hi,

I have recently setup monitoring on our company's website and i am having a few problems with the way the .Net agents are reported

The setup is as follows:

6 hosts run both, IIS Web Services and .Net, to serve and render the page for visitors.

I have setup webserver and .net agents and traffic is being tracked correctly.

The issue i am having however is as you can see below, the .net agents are listed multiple times with [1] after their agent names, the causes dynatrace to report "Application Process Unavailable (unexpected)" and these are then raised as incidents in my incident dashboard

I have looked into this a little further and it seem like the process is ended then starts striaght back up again then getting listed as [1] the next time this happens, the [1] agent is disconnect and the original is re-instated

This can be seen in the host dashlet below:

As far as i can see there is no disruption in service so its not a massive issue, but i would like to understand it more and hopefully prevent the incident being raised and reported across the system

Any help would be greatly appreciated

Thank you

Joe

5 REPLIES 5

Babar_Qayyum
DynaMight Leader
DynaMight Leader

Hello Joe,

I would suggest you to check the following things for more understating:

  • Agent Logs
  • Collector Logs (Also the Collector Sizing dashboard for any Buffer Saturation, CPU Usage and Memory Usage)
  • Check the instrumentation repose time from the 'Agents Overview' dashboard. Just add the below columns from the header with the right click:
    • Average Instrumentor Response Time
    • Maximum Instrumentor Response Time

Regards,

Babar

kristof_renders
Dynatrace Pro
Dynatrace Pro


Hi Joe,


This behavior is caused because AppMon adds an index in cases where the agent comes back sooner than AppMon can unregister it.


This behavior can be changed, however by adding a debug flag


Please check this link in the documentation and let me know if this resolves the issue.


KR,

Kristof

joe_wells
Participant

Hi @Kristof R. & @Babar Q.

Thank you both for your quick responses, i have looked in to both of these options.

Were running 6.3 here so i have referred to this older documentation: https://community.dynatrace.com/community/display/...

I will give this setting change a go, however, do either of you know what the number is at the end of the agent name, after the :, is it a process ID, or port number, will changing this setting enable them to be listed as the same agent with different numbers?

The collector logs show entries for the agents being in a disconnected state and the reason code is listed as "Connection reset"

The agent logs offer no really data on why the process may have been reset, in response to the above question i think the numbers refer to the process ID as the logs are listed like this too, looking at a log of the disconnected process id, there is nothing other then the first start information

Here is an image of the avg and max response times

Hello Joe,

You are absolutely right. The PID is part of every agent name.

Regards,

Babar

joe_wells
Participant

If it helps at all, i have discovered that the incidents that are being logged appear to be on some type of schedule

Any idea how i can extend this window to see these incidents over the larger number of days?