Uneven load balancing when forwarding logs to syslog server using OpenShift Logging

Issue

• Delay in log ingestion by external syslog servers which are backed by a Load Balancer, when forwarding logs using Red Hat OpenShift Logging.
• When forwarding logs to syslog endpoint, the load is not evenly distributed to syslog servers which are behind a load balancer.
• Vector opens a single TCP connection to a Load Balancer, the load balancer typically sticks to the same backend for the duration of that connection, until the pods are restarted.
• How to configure max_connection_duration_secs for Vector socket sink ?