Multiple workflows from different SCOM enviroenment.

Question

Hello All,

Thanks in advance,

Currently we are in the process of SCOM LCM. We have multiple servers are reporting to SCOM 2016 (Prod/Test) SCOM 2019 (Prod/Test).
We have multiple script based custom MP and heavy MS MP’s like SQL . Since the servers are reporting to all four environments. These MP’s will run multiple workflows on servers that may create performance issues in servers.
My doubt would be is there any self-optimization in SCOM agents if the same workflows are running from different environments.
I believe Alert cook down functionality will work only for same environment.
Can anyone help me on this.

Thanks & Regards,
Anu

Answer 1

Hi @ANU ,

You are also correct about Cookdown - it does not work for the different enviroments, it can only work for one individual MG. In order for it to work, the same parameters need to be passed to the data source, which is of course impossible for different env. Thiss also explained here:

Operations Manager Management Pack Authoring - Cookdown
https://social.technet.microsoft.com/wiki/contents/articles/15218.operations-manager-management-pack-authoring-cookdown.aspx#Criteria_for_Cookdown

Whaat George mentioned above is also true - if the Health Service reaches certian threshold it will restart automatically in order to free up resoruces. This was a known issue on SCOM 2016 and I alslo blogged about it:

SCOM 2016 – Agent (Health Service) high CPU utilization and service restart
https://social.technet.microsoft.com/wiki/contents/articles/52853.scom-2016-agent-health-service-high-cpu-utilization-and-service-restart.aspx

what you can do to prevent this if your agents are heavily loaded and are restarting is to increase the agent thresholds even further:

Health Service Handle Count Threshold
Health Service Private Bytes Threshold
Monitoring Host Handle Count Threshold
Monitoring Host Private Bytes Threshold

Please see the refence to Kevin Holmans blog post in the article I wrote.

Besides this there are no other known way of doing perfomance optimizations on your aagents. This is because their performance highlyy depends on how all the MPs are written (especially the custom ones) and how Targeting and Overrides are handled in your environment.

I hope I could clarify this for you!

----------

(If the reply was helpful please don't forget to upvote and/or accept as answer, thank you)
Regards
Stoyan Chalakov

Answer 2

Hello @ANU ,

You're correct, a multihomed SCOM Agent to multiple Management Groups where the same workflows are configured, will execute multiple instances of those workflows (one for each instance / Management Group).

Cook down works only in the same Management Group.

When you are in migration processes, requiring your agent to be multihomed, then it is recommended to mind the CPU / Memory usage of the SCOM Agent and ensure that your healthservice and monitoringhost memory usage monitors are not hitting the thresholds (generating Agent restarts).

BR,
George