Don't reload when use-cluster-ip endpoints update, and change the ingress use-cluster-ip implementation to use the cluster ip instead of the fqdn
#588
| Job | Run time |
|---|---|
| 11s | |
| 10s | |
| 12s | |
| 42s | |
| 7s | |
| 1m 22s |