I have CRDB cluster fronted by haproxy installed based on documentation. haproxy config created by crdb tool. connecting to cluster nodes directly is always ok and permanent. connecting to cluster nodes via load balancer drops frequently. I have to write a loop in server to ping db constantly to keep connection alive. Otherwise always first attempt to access db comes back with an error like ‘driver: bad connection’ and this is because haproxy dropping connection. I’m also using this in haproxy config:
option pgsql-check user root
Am I missing anything in my setup?
Any help will be appreciated.