May I ask you some questions about Dynatrace baseline?
Customer has monthly-cycle business processing like closing day on the end of the month.
I am wondering if Dynatrace baseline can recognize monthly peak/monthly pattern by application traffic.
I have read the help document as below.
According to it, Dynatrace baseline learns daily and weekly patterns. There is no mention about monthly pattern.
So, Dynatrace baseline cannot learn monthly peak/monthly pattern.
Is my understanding correct?
I summarized my understanding about baseline of regularly monthly peak in PowerPoint. So would you please refer to it?
And I would like to ask you questions.
In the PowerPoint, which Case can we implement with the current Dynatrace, Case1 or Case2?
(We are very happy if we can implement Case2.)
Is there any special way to achieve the Case2?
Finally, please let me add one more question.
Have you ever received a request from customers that they want to recognize monthly peak/monthly pattern as a normal event? And if you have, how did you deal with it?
Right, the Dynatrace automatic baseline detects abnormal behavior on a daily and weekly seasonality. We do not support monthly or yearly seasons.
Reason for not supporting a yearly pattern is that you would need multiple years learning times, same as for monthly seasons.
Most businesses show a clear daily and weekly business cycle anyway, so we chose those as the default for our alerting.
Sometimes we get the request for detecting 'black Friday' yearly patterns, but that's in most cases not possible as the underlying infrastructure as well as the software changes within a years timeframe too much to have a reasonable learning period.
Thank you for response! The information you gave me is very helpful. I understood why Dynatrace does not support monthly or yearly patterns.
But my customer would like to detect monthly patterns.(He would give up yearly patterns.)
Is there any workaround for detecting monthly patterns?