cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Knowing AI behaviour in different scenarios

elena_renedo1
Participant

Hello,


We want to know how is the behaviour AI (Artificial Intelligence) in the following scenarios:

  • Events that happens every weeks. Ex.: Every Monday at 4:00 a.m. a process backup is launched.

What was the behaviour of the AI? From what week do you consider it a frequent issue? Maybe the second week the event was launched? What happened if the event is launched alternate weeks?


  • Public holidays

What was the behaviour in a public holiday happened in a working day? Does the AI save days from previous years? And what happen if not every year that day is a public holiday?


  • Commercial campaigns that a weekend

At what point would the AI begin to realize that the increases in requests, degradation of response times, are due to a specific scenario and not to a problem?


  • Punctual day of massive load

For example, a Black Friday day that it can be at the same time than the previous day or not. Does it keep any kind of history? Does this history depend on the number of transactions that are being stored?


Regards,

Elena.


1 REPLY 1

In general dynatrace is learning baselines in daily basis for whole week. If some incident is every week on the same time AI should stop alerting after some time. For other cases you mentioned that are incidental it is possible that DT will alert you about traffic to low or to high (depends on scenario). You can’t do much in this area right now.

For scenarios like deployments / backups etc you can always use maintenance window for suppressing alerts during planned work.

https://www.dynatrace.com/news/blog/manage-planned-downtimes-maintenance-windows/

Sebastian