I have a couchbase cluster with 4 hosts (4 CPU, 32 GB ram).
This cluster is used by thousands processes which get data from it. I have 8K permanent connections on the cluster (data from couchbase ui).
The bandwidth uses by couchbase is 50MB/s per hosts to send data (don't seem very heavy).
But I get client side timeout. About 4 per seconds and I don't understand why.
I set somaxconn and tcp_syn_backlog to 2048 per hosts but it doesn't change something.
Do you have any idea ?
Edit from comments:
I have this kind of logs from my client:
Error from Couchbase, passing in non-blocking mode (EXC: Key=u'fieldmetadata', RC=0x17[Client-Side timeout exceeded for operation. Inspect network conditions or increase the timeout], Operational Error, Results=1, C Source=(src/multiresult.c,316), Tracing Output={"fieldmetadata": {"i": 15890195098182236653, "s": "kv:Unknown", "b": "MetaDataV3", "r": "ltg-aus-couchbase-3:11210", "t": 2500000}}>)
I have 3 nodes in my cluster. All clients and cluster is on Azure VM.