I have an interesting observation with the grpc async server. So, I have 2 worker threads, A and B, each one has its own queue. I ran a stress test, initially I see that both the worker threads are serving to clients. After some time, I don't see any activity from thread A. A thread is not stalled on anything, it is waiting on conditional queue for the event. This thread has already scheduled an event in condition queue to accept new connection. I see that for around 4 hours this thread was not scheduled. Now, other thread B, gets stalled in an application after 4 hours. After this point if I try to connect new client to service, I don't see this request being served. So, it seems like all those requests went to condition queue for B where my A is free. This brings up the question that how is load balancing done internally for the requests?
-- You received this message because you are subscribed to the Google Groups "grpc.io" group. To unsubscribe from this group and stop receiving emails from it, send an email to grpc-io+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/grpc-io/6ce16c81-469f-4204-a181-841e1388fbc3n%40googlegroups.com.