RFR: 8157023: Integrate NMT with JFR
Erik Gahlin
egahlin at openjdk.org
Thu Dec 1 13:09:11 UTC 2022
On Thu, 1 Dec 2022 12:32:12 GMT, Stefan Johansson <sjohanss at openjdk.org> wrote:
>> src/hotspot/share/services/memReporter.cpp line 861:
>>
>>> 859:
>>> 860: MemBaseline usage;
>>> 861: usage.baseline(true);
>>
>> Perhaps we could collect event timing here, rather than reading the timer for each memory category? That way it’s easier to align events from a single snapshot.
>>
>> for this sort of event, I’m not sure if it’s preferable to declare only a start time, start==end, or measure start and end time around the ‘baseline(true)’ call.
>
> Sounds like a good idea. The only potential problem would be a thread view trying to visualize all event when they were sent and seeing a lot being exactly equal might be strange. I'll consult the JFR guys to see what's best to make it easy to verify which samples are from the same baseline.
Sounds good. We have tried to use the same timestamp for periodic events that come together. This simplifies processing of events, since the timestamp can be used as a key.
Looking at other periodic events, they only set period="everyChunk", which I think means startTime=false implicitly. I think it would be good if this follows the pattern.
-------------
PR: https://git.openjdk.org/jdk/pull/11449
More information about the hotspot-dev
mailing list