By what kind I meant what latency do you see eg. when you do a simple ping.
Interesting lines in the log: WARNING: DNS Reverse Lookup of 172.31.8.43 is slow, took 4525ms
I bet your client machine is a Windows one? Each IP in the cluster is reverse looked up so that we have the hostname for various reasons (one of which is meaningful logging), but we noticed that often on a windows host the operation’s performance is very poor.
Here it takes 4.5 seconds for 1 IP, which adds up… Can you maybe try to configure a hostfile so that kind of resolution becomes trivial for windows? (not sure how to do that in that environment)
That said the timeouts you set are huge and 6 nodes reverse lookup shouldn’t cause a timeout with that kind of duration…