Document overload manager configuration

I am realizing that I depend on an undocumented behavior. The overload manager triggers when memory usage reaches 80% and stops accepting requests:

https://github.com/envoyproxy/gateway/blob/8e813bdd05ecbccf0650c15f9b309b4ca49384e0/internal/infrastructure/kubernetes/proxy/resource.go#L477

https://github.com/envoyproxy/gateway/blob/8e813bdd05ecbccf0650c15f9b309b4ca49384e0/internal/xds/bootstrap/bootstrap.yaml.tpl#L322-L345

I have configured the HorizontalPodAutoscaler threshold to a lower value, so I would consider any change that lower the overload manager threshold below it, a breaking change.

Maybe a new "Overload Manager" page under Traffic? Although it is not really a "task" since it is not configurable at the moment (#3604)