Dynatrace took 6-8 minutes to create the problem after it happened. we tested this by shutting down JVM's . But once problem is generated, problem contains exact time when the issue was happened. it will be great, if dynatrace able to generate the problem a little bit more quicker
process group was stopped at 14:05
Problem #862 in dynatrace was generated at 14:11
Email alert came at 14:12
Ticket in Netcool queue 14:12
As I remember time to alert after process crash is obligatory and cannot be changed. This is because of preventing falls positives when for example iis pool is recycled.
You can post RFE for this idea to have option of overwriting this time.
I believe this is due to how the problem detection actually works (correlation with other event, checking of dependencies, etc...). Even for external events from 3rd parties it is few minutes (3-4 mins based on my observation) until the problem is generated within Dynatrace.
I don't think it can be much faster. However, the problem always contains the exact time when the issue was detected.