After upgrading from OpenShift 4.10.17 to 4.10.25 customer hitting several failed liveness and readiness probes

4.10.25

Not reproducible on other cluster, only one of many cluster is affected

1.
2.
3.

Customer is hitting high connection time in liveness and readiness probes which causing NotReady Pods and redeployments.

oc rsh prometheus-k8s-0 curl -w @- -o /dev/null -s "http://localhost:9090/-/ready" <<'EOF'
>     time_namelookup:  %{time_namelookup}\n
>        time_connect:  %{time_connect}\n
>     time_appconnect:  %{time_appconnect}\n
>    time_pretransfer:  %{time_pretransfer}\n
>       time_edirect:  %{time_edirect}\n
>  time_starttransfer:  %{time_starttransfer}\n
>                     ----------\n
>          time_total:  %{time_total}\n
> EOF
    time_namelookup:  0.078855
       time_connect:  7.181930
    time_appconnect:  0.000000
   time_pretransfer:  7.181973
      time_edirect:  0.000000
 time_starttransfer:  7.182554
                    ----------
         time_total:  7.182616

No delay in connection to loopback interface

prometheus dump, must-gather, sosreports and further metrics attached to support case