Currently, occasionally an occurs that "incident" can not be detected.
The setting is as follows.
Evaluation Timeframe: 10 Sec
Measure: BT (Threshold:Upper Severe: 7000 ms, Upper Warning: 2000 ms)
Although it is not every time, response time exceeding Upper Severe Even if it is measured, it may not be detected as an "incident".
Is something wrong to satisfy the above conditions?
Also,Please tell me what each of the following items behave.
For example,What is the difference between "Evaluation timeframe: 1 minute - Aggregate: maximum" and "Evaluation timeframe: 1 minute - Aggregate: last".
I saw the manual, but there are many unknown things,So I do not know how to set it up.
Evaluation Timeframe = the timeframe for which we're looking at data. For example, if the evaluation timeframe is 1 minute, we're using a 1 minute rolling window for analysing data against your measure, threshold, and aggregation.
Aggregation = how do we want to calculate the data. For example my transactions gets 20 hits in the incident timeframe and I'm setting the incident on response time - do I want just the last value of those 20 hits, do I want the average, or maybe just the maximum value for those 20 hits, etc.
Threshold = which threshold should this incident alert against? (warning level, severe level, both?)
Evaluation Timeframe 1 minute + Aggregation Maximum + Severe Threshold (7000 ms) -> means that this incident will trigger if in any 1 minute the maximum value of my transaction's response time goes over 7 sec.
Evaluation Timeframe 1 minute + Aggregation Average + Severe Threshold (7000 ms) -> means that this incident will trigger if in any 1 minute the average response time of my transaction goes over 7 sec.