We are running a 36 node storage cluster with 9 proxies. The application written in C++ starts to get timeouts from Coherence extend which we have set to 5s around 30 minutes into heavy put load. At this point it will fall over as it continues to fall behind. The problem came about when we updated from 18.104.22.168 -> 22.214.171.124. Is there any known issue with the current release which would trigger this behavior?
I have also noticed at some points the client will appear to stop functioning and we will no longer get timeouts.