This message was deleted.
# develocity
s
This message was deleted.
n
This is unfortunately all the information that the JVM exposes. Can you tell us your use case here?
r
The usecase would be to optimize the memory size for each generation in the case of some are being suffocated with the defaults, so you can optimize the overall performance. In our case we have enough slack to just increase the heap and do not even need to go to that level of fine-tuning yet. So this is not something we need to look into right now. But since it was reported for other GCs I was curious to understand why it was missing here.
n
I think a memory profiling tool is more appropriate here. Do you know how to obtain this information for the G1 gc from the JVM?
r
I think a memory profiling tool is more appropriate here.
Yes. But we are not constantly profiling the builds. The way we use GE is to spot these “red flags” in the builds, so if we identify slow builds and see that memory peak is at 100%, would be a reason to start profiling. And the peak might be just in one of the other generations, not necessarily in the old. 🙂
Do you know how to obtain this information for the G1 gc from the JVM?
I honestly have no idea about how to get this. I know is available, because if I open a visual memory profiler locally for a build running with G1 I can see the information of each generation. But I am not familiar with how the gradle enterprise plugin is retrieving this from the process.
Actually, I don’t believe this is reporting just the old gen, but the whole heap usage (because the 8gb there is the max heap we declare). It’s kind that is not differentiating generations and, but then the actual label is wrong