22 May 2021 03:12 AM - last edited on 24 May 2021 08:21 AM by Karolina_Linda
All of our Linux hosts run a process named Chef (chef-client to be precise). This process runs every 30 minutes. As a result, all day, every day, every 30 minutes, on every Linux machine (we have 100's) I get an Event 'chef-client restarted'.
These machine are all monitored as InfraOnly.
Is there any way to tell Dynatrace OneAgent to NOT monitor this specific process or stop it from raising an event for this process? If there is an option do I have to do it for each host or can I make this change tenant wide?
22 May 2021 08:43 AM
Hi @ct_27
You can do that from Setting --> Processes and containers --> Process group monitoring -->
HTH
Yos
07 Jun 2021 10:51 AM
That only affects deep monitoring... not the basic infra level monitoring of processes, or the event creation regarding process restarts.
07 Jun 2021 08:17 AM
Thank you for your suggestion. We had tried it right away but did not stop the Events. We tried many other flavors of the same solution, even with our Guardian's help, but could not get the Events to stop.
The process itself is scheduled using systemd timer, so we're not sure if that has to do with it and why Dynatrace seem unable to identify and ignore it.
We have a similar process on all of our Windows machines (the above chef is for all Linux machines) named MonitoringHost.exe which is related to SCCM. This one also raises many Events we have been trying to get Dynatrace to ignore. No luck with this either.
The problem is that each of these raise one event every 30 minutes. So, 48 a day, 144 in 3 days. This amount of noise can make it really difficult to find important Events of similar type. We've missed major process restart Events for important processes because the 1 was mixed in with 48 other ones.
Any additional suggestions or ideas are welcome. We're trying to get Dynatrace to stop raising Events on chef-client (https://medium.com/@tharmini7/chef-client-2eab9a45f10d) and MonitoringHost.exe
08 Jun 2021 09:00 PM
You might want to open a support case on this. For the time being, you can add a Maintenance Window to suppress that process as needed to ignore and not alert.
09 Jun 2021 09:14 PM
Chad,
Thank you so much for your suggestion. I created a maintenance window and told DT to not create problems and not alert but it still recorded the Events. I spoke to our system Admin who mentioned as well that this job is run through Systemd Timer (more modern solution over cron).
This technical difference really shouldn't make a difference but still worth noting.
The screenshot below shows that although the process is in maintenance the Events are still getting created.
I'll go ahead and create a support ticket.
24 Jun 2022 06:41 AM
Hy CT,
Any luck with that MonitoringHost.exe process monitoring? Have you managed to stop Dynatrace from generating events related to it?
Br,
Sorin
24 Jun 2022 07:39 AM
Yep, after more back and forth with support we managed to get a custom black-list entry added to our tenant. It was the only way to stop the events.
It's so nice to not have 3365 Process restarts every 30 days. We can actually see real Events now that actually have value.
24 Jun 2022 07:45 AM - edited 24 Jun 2022 07:46 AM
That's great news! Can you please give me the support ticket number? I'd like to open one also and have that black-list entry added to all our customers. I'll use your ticket as a reference.
Cheers,
Sorin
24 Jun 2022 07:49 AM - edited 24 Jun 2022 07:55 AM
#32106
" opening up a new support ticket for the specific process and request to have it put in our process blocklist."
27 Jun 2022 08:31 AM
@ct_27 glad to know that your query was solved by support team. It would be helpful if you can document the exact steps and post it in some kind of blog in the community forum.