Larger than expected http connection counts observed when using some network load balancers

Issue

In some environments running with a Network Load Balancer in front of Redpanda,  it was observed there were a higher number of http connections to Redpanda (HTTP Proxy/schema registry etc) than expected per metric  vectorized_httpd_connections_current
 

Detail 

Investigations showed that the http connections to Redpanda were being silently dropped by some Network Load balancers but never being cleaned up on the Redpanda side 

A fix was created for Redpanda to implement TCP keepalive on the Redpanda HTTP API's
 

Resolution 

The keep-alive fix was implemented via  PR: http: Add TCP keepalive to http APIs #28838
which is  available in Redpanda Release  25.3.2  and higher