Utilize the "Events API"

[cmu]

We would like to use the Events API available to create problems in case of issues with a flow in one of our systems. It will be events if the flow fails for any reason.

We plan to send events from one application whenever an error is returned in the flow. Could be exception, no response or other issues.

We have not used this API before, and we have a couple questions that I hope you can answer.

If we post multiple events, will Dynatrace then, through AI, know when to raise a problem? If we post events as problems, will the problem events somehow be grouped/collected into one, if they are similar? Or will it be 1 problem pr. event?

Is there a limit to the number of events pr. minute or so?

I'll most likely have more questions later, but hope someone can help me out.

[skrystosik]

Such static events post via API are not considered by Davis. They will be visible as single instance problems I think. Davis can correlate events related to monitored entities, but they will be visible as potential root cause when this entitiy will be involved into problem.

Sebastian

[JamesKitson · Accepted solution]

The new version of the AI engine (Davis) does indeed ingest and analyze third party events:

https://www.dynatrace.com/support/help/shortlink/next-gen-ai-analytics#third-party-event-ingests

A lot goes into it but keeping it simple: when you post events they will be "attached" to various entities in Dynatrace (services, process groups, hosts, etc...) and so if there are dependencies or interactions between the entities you're attaching events to you'll likely have them grouped into one problem.

If they are standalone events for instance on process groups that don't interact with each other then you'll likely end up with multiple problems. The type of event isn't what is used when grouping them all together, it's how everything is mapped in that smartscape.