Does Helidon have a request queue like Jetty (presumably) does? I wonder if Helidon is more susceptible to burstiness, whereas Jetty's queue keeps the throughput more consistent.
I ask because the only built-in virtual thread executor has no thread cap.
Andrew O'Hara
01/03/2024, 9:26 PM
Moving to a thread...
I ask because because this service has a pretty sawtooth steady state usage.
j
James Richardson
01/03/2024, 9:44 PM
Hmm. Could be that your service itself uses a bit of memory and the number of simultaneous requests being handled blows the limit?
James Richardson
01/03/2024, 9:45 PM
So memory used not by loom stuff, but by things handling requests...