I am trying to figuring out what is the best approach to follow when it comes to analyse memory leaks with Dynatrace.
Following the suggestions below:
It makes sense to manually trigger memory dumps in non-prod envs, where you have well defined load test scenario and you can control and schedule dumps in the proper way.
what if we get an unexpected memory leak in production environment? In that case, automatically triggered memory dumps are critical in order to have some data to asnalyze but it seems Dynatrace is not offering this kind ok capability.
The only valuable option that I see for this scenario is adopting the "Continuous memory profiling" in which we have the memory consumption in real time. Unfortunately, this is only available for java11+ and most of our systems are <11 or based on different technology.
What are the best practice in this case? How can we use dynatrace the best way possible?
In general there is no possibility to automaticly trigger memory dump. For now there are only such options like that you've mentioned. There are no plans (according to my knowledge) to allow triggering of memory dumps via API so in such case you can only do it reactive if you cannot use memory profiler.