pgcat pods running on k8s dropping connections while rollout

[asaray11]

Is your feature request related to a problem? Please describe.
I use pgcat pods for my Postgresql connection pooler in k8s. I deployed pgcat with helm using this repo. Since autoreloader not working correctly in k8s, I run 2 replicas of pgcat. Thus service will not be interrupted in case of config changes. But when I make tests I saw connection problems occurred during rollout of pods.

This is test query:
root@psql-client:/# PGPASSWORD=xxx bash -c 'while true; do echo -n "$(date "+%H:%M:%S") -> "; if ! psql -h pgcat-helm.pgcat.svc.cluster.local -p 6432 -U xxx -d xxx-t -c "SELECT now();" >/dev/null 2>&1; then echo "Connection error!"; else echo "OK"; fi; sleep 0.1; done'

When I run this test and make a rollout it gives connection error;
05:55:04 -> OK
05:55:04 -> OK
05:55:04 -> Connection error!
05:55:04 -> OK
05:55:05 -> OK

Describe the solution you'd like
I forked repo and changed deployment.yaml and added "initialDelaySeconds: 5" then it worked without error.

      livenessProbe:
        tcpSocket:
          port: pgcat
        initialDelaySeconds: 5
      readinessProbe:
        tcpSocket:
          port: pgcat
        initialDelaySeconds: 5

Describe alternatives you've considered
I will be glad to hear if anybody can run pgcat autoreloader in k8s.

[asaray11]

Opened pull request #940