topic Re: Why are the problems open for more than 12 hours? in Alerting

Why are the problems open for more than 12 hours?

Yssa — Thu, 10 Feb 2022 10:06:52 GMT

Hi, could you please help me with the following;

I have read that Dynatrace automatically closes Problems that are open for 12 hours. But, I have more than one event of type Resource for example, for "Disk low" with more than 51 days open.

Can you explain me how the automatic closure is applied for each type of event?,Why do I have open events with more than 12 hours and they are not automatically closed?.

Thanks

Re: Why are the problems open for more than 12 hours?

r_weber — Thu, 10 Feb 2022 10:59:16 GMT

Hi @Yssa ,

in case of a disk low scenario, this is based on a static threshold of (default) 97%. So as long as your disk is above this threshold the problem will stay open as the condition for the problem is constantly met.

kr,

Reinhard

Re: Why are the problems open for more than 12 hours?

Yssa — Fri, 11 Feb 2022 12:52:02 GMT

Hi @r_weber , thanks for answer,

So, the automatic closes of Problems doesn't apply to events Type: Error events, slowdown events, resources events and custom alerts?.

Regards.,

Re: Why are the problems open for more than 12 hours?

r_weber — Fri, 11 Feb 2022 20:32:06 GMT

Hi @Yssa,

no the DAVIS AI analyzes the circumstances, causation and correlation of events and combines those into problems. So one problem might be the result of one or multiple events.

Multidimensional baselines on metrics is also a contributing factor. Davis continuously evaluates data and decides based on that if a problem is resolved or not.
For baselines (e.g. a simple response time violation) these are constantly updated. So if your response time behavior of a service changes due to a deployment, let's say the median moves from 500ms to 700ms you will for sure get a problem (slowdown) but if the baseline adapts (it is continuously updated) you will have a new "normal" and the problem for 700ms will eventually close - simply spoken.

There is a lot of information here.

kr,
Reinhard