15 Oct 2020 09:13 PM
Hello,
Is there any way to prevent raising problems which are already known?
For example specific "User action response time degradation" problems are being raised multiple times per week. However they are already investigated and corrective actions will be taken in the next three months.
Is there any way to prevent such problem from reappearing again until corrective action is in place?
Thank you,
Martin
Solved! Go to Solution.
16 Oct 2020 08:43 AM
Usually, unless you have turned off the request issue detection in tenant settings, Davis would pick it up and won't alert you unless the problem severity increases.
If this does not happen for you, then probably those problems differ. For example it is for a different user action. You might want to finetune your settings - in this example setting appropriate expected values for user action metric.
Also when considering user actions, and this is a common issue, you don't have user action rules precisely defined. That means, different user activities result into the same user action name. Then you might experience high volatility values for the key user action metric and noisy alerting happens.
16 Oct 2020 11:46 AM
Hi Julius,
We have the action names set correctly and unique.
Davis is reporting response time degradation for example from 1sec to 5 sec for some actions and probably they are not so repeating to consider them as frequent issue. Still they keep coming and users are being alerted on something that is already reviewed. The Apdex thresholds are large enough and this response time degradation will not reflect on Apdex metric.
What is the expected value for user action metric that you are referring to?
Anyway to control the thresholds for user action response time degradation?
Thanks,
Martin
16 Oct 2020 02:54 PM
Sorry, maybe I was not exact in my answer. You should finetune the anomaly detection on those actions (must be key user actions), not the apdex.