Allow envoy StatsMatcher to be configurable

[davidgiga1993]

Envoy by default exposes tons of metrics, a lot of them are on a per-cluster basis, causing the number of metrics to increase with the number of ingress endpoints. This can cause issues with scaling Prometheus as well as unnecessary network traffic for scraping.

Other projects are facing similar issues (https://github.com/hashicorp/consul/issues/12509)

Therefore I propose to allow the envoy StatsMatcher to be configurable so we can choose what metrics should be excluded directly at envoy.