Hello,

I can only find https://support.microfocus.com/kb/kmdoc.php?id=KM02747433&fileName=PerfSizingGuide_OMi10.60_PE10.60_A.01.07_CustomersPartners.pdf, but I'm sure there is a newer version somewhere.  I think this is difficult question to answer, because it depends on your monitoring (events per agent/per day) and questions olike how much health checking you choose to make.

Containerised OBM:
docs.microfocus.com/.../Sizingdepl

Classic OBM:
https://docs.microfocus.com/doc/Operations_Bridge_Manager/24.2/ObmDeployment

I think this table from https://docs.microfocus.com/idoc/Operations_Bridge_Manager/24.2/Install/OBMInstall-245126 gives you an idea of OBM sizing.   This describes the "Extra Large" setting for more than 15000 nodes.  I would expect extra large sizing to suffice:

Processor. The following table lists the CPU requirements. We recommend using 2.4 GHz CPU cores or faster.

In a virtual environment, make sure the number of virtual CPUs is equal to the number of physical CPU cores used.

Dear Duncan ;

thanks for your reply and the document but it is mentioned here that OBM can handle upto 336 event/sec but is that number means the actual fault detected events or also included informational data came from the Agents 

as if it is included also the actual data this will be very low rate as we expect millions/sec of streaming events including informational data

what do you mean "included additional informational data" that comes from Agents??

OA only sends events to OBM when a fault is detected, and that's the number of events per second mentioned above.

I have no idea how you are getting "millions/sec" with about 13K Agents. please clarify.

One last thing to mention - event processing performance will be vary depends on what additional processing YOUR OBM is doing (for example: EPI scripts, SBEC rules, invocation of Runbooks or Ticket integration, etc....).

hello asaf thanks for your reply ,  do you know a valid way to calculate how many event/sec i am receiving or expected to be recieved on a logic equation 

Hello,

One way - find out when the process started:

# ps -ef |grep opr-backend |cut -c -100
root 11459 3227 0 Mar06 ? 00:00:00 /bin/sh /opt/HP/BSM/bin/opr-backend_run.sh
root 11467 11460 0 Mar06 ? 08:06:56 /opt/HP/BSM/JRE/bin/hpbsm_opr-backend -DskipDDPKIP
root 76288 6900 0 14:55 pts/0 00:00:00 grep --color=auto opr-backend

So that's the 6th March 2024.

Then run opr-jmsUtil.sh looking for opr_gateway_queue.  

# /opt/HP/BSM/opr/support/opr-jmsUtil.sh

Here you can see that there have been 87440 events sent since opr-backend started - which is 6th March.  Since today is the 23rd May, so that's a that's 78 days.  Therefore, this OBM server has received 87440/78=1121 events/day.  There are 86400 seconds in 24 hours so 1121/86400=a very small number.  It will be different on your system.

Following up on Asaf's points, OBM can deal with 1000's of events, and not all of them end up being used.  If you have a business case to manage 12000 nodes and you have OBM to manage it, then please set the system to "Extra Large" and test.  I'm constantly amazed about how many events/second OBM can process when well configured :-)

Hope this helps.