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

Detection of Problems when data is reset with Anomaly detection

knakayama
Contributor

Hello,

I have a question about Anomaly detection.

When you press Reset in "Reference period", including the day you clicked,

I think that seven days of data will be reset.


(For example, if you reset it at 9:00 on February 26th,

Data of 9:00 on February 19th to 9:00 on February 26th is reset. )

This is an effective means when you know that the trend of the system is different from usual, such as temporary work.

Do you have my knowledge, such as the period of reset?

Based on the above, there are things I do not understand.

For example, suppose you pressed the reset button at 9:00 on February 26th.

On 28th February, when the traffic was low unlike usual,

I think the data on February 21th is reset,

Is it detected as Problems compared with data of more than one week ago, such as February 7th and 14th?

Do you know anything?

I am looking forward to your reply.


2 REPLIES 2

wolfgang_beer
Dynatrace Champion
Dynatrace Champion

If you click on rest baseline the automatic baseline detection ignores the past week of reference data and starts over to collect a new reference. For load baseline this means 1 week of learning again before alerting is active again, for performance and error baseline this means 3-5 hours learning time before alerting is active again.


In staging envs where most of the time there is no traffic, and when there is load test, the performance might be bad, in this case, how is new baselines established after reset?