Event management limitation

[Jakub Drd_k]

Hi, we are testing event management on our dev instnce and creating events via rest api. But we able to create only cca 250 events per minute and we need create 300 events per second. Is there any limitation or can you give us some lead? Can we create 60 parallel sessions to reach what we need?

[Scheuerman1]

What happens when you create 251? Seems to me that you might be hitting a limit on the API side of things.

[Tony Branton · Accepted solution]

Event Management can process around 400 events per second on a Production instance configured to use all application nodes and with multiple scheduled workers (e.g. 2 per node).   This is configured in the Event Management Properties page.

Non-Production (i.e. Dev) instances only have a single application node, so the only performance tuning you're able to do is increase the number of scheduled workers   (e.g. to 4) however - at best - this will allow you to process around 100 events per second.

Non-Production instances are useful for configuring and testing event rules, but should not be used to evaluate event processing performance as the instance architecture is not equivalent to a Production instance.

[roy_silon]

Taking the opportunity to update this quite old resolution with updated Scale numbers for ITOM and TSOM (updated to Xanadau Family Release):

Event Management - 3k-10k events a sec, with 1 million services which 20% of them getting alerts.

HLA – Standard deployment with single MID 18K logs/sec with multiple MIDs 50K logs/sec; We can scale to 240K logs/sec. 

Customers who are interested in more can approach us and receive a more detailed white paper with benchmarks and testing methods.

Mertic Intelligence -  7M metrics per minute.

[srajput]

How do we get access to the detailed white paper?