I have an incident with threshold of 8 sec and i want to get the notification once the average response time is greater than 8 sec. I have aggregated the purepath by Average but still i am getting the notification if only once its breaching the threshold. Can we aggregate the purepath by Mean?
I want to exclude first highest and lowest occurrence of purepath form the purepath list. If only once its breaching the threshold i am fine with that. I should get notified only if its occurring more than once
Appreciate your quick response.
Hi Bharti G.,
I can see a couple of approaches to achieving the desired outcome that you would like to achieve.
The first approach would be to add another condition to the incident rule settings. I suggest that this condition is based on a count measure of the number of occurrences with a threshold of two (2). The existing 8 sec average threshold is also a condition in logical AND with the other condition.
The second approach could be to utilise the Rate (Calculation) measure against your base measure. Then creating an incident rule based on that rate measure and some threshold that you determine would be appropriate. This is only slightly more trickier than simply adding another condition to your existing incident rule.
In either case, you would need more than a single measure to accomplish your desired outcome.
As an additional note, you may also want to consider the evaluation timeframe value as the default will be only 10 seconds. It may be more valuable for you to increase this.
AFA my experience, I would not recommended to send alert based on PUREPATH response time. Instead you should always group relevant purepaths by defining business transactions and then create an incident which two measures in it:
1. BT with Response time threshold say Upper Sever: 6000 ms
2. Count Threashold metric containing above BT/M with upper sever count as say 200.
This will effectively mean alert when there are 200 purepaths of that transaction having avergae RT of 6 sec or more. Which is better.
Here is a snapshot how you can do it:
Even better approach is to include your BT for the Auto-Baseline subscription. However do not overload many BTs into auto-baseline.
Also, to exclude noise the Percentile is better filter/technique for your case. Look here.
Let me know if you have specific requirements.
Award best answer if helpful 🙂
@Rajesh S. Thanks a lot for your response. Exactly i was doing the same what you have explained but the issue is still i am not getting alert notification for average response time, instead of that i am getting the alert notification even if purepath is breaching the threshold just once within given evaluation timeframe. Please find the given snapshot for reference.
Apprecaite your help.