Commit Message: admin: support /drain_listeners?skip_exit
Additional Description:
Due to changes in Kubernetes (https://kubernetes.io/blog/2023/08/25/native-sidecar-containers/) around sidecars, there is a more optimal way to manage the lifecycle of Envoy as a sidecar in Kubernetes. However, it is not currently possible with the current knobs.

Our goal:

1. When the pod starts terminating, start "draining" inbound connections. Accept new ones, but respond with GOAWAY/connection=close to discourage clients
2. In parallel (outside of envoy), clients should be discourage to making new connections generally
3. Once the application container shuts down, we have nothing left to do; as a sidecar proxy, we exist only to serve the application container. Shutdown ~immediately

The time between (1) and (3) can be quite large (but not unbounded; Kubernetes has a max time before it will SIGKILL everything).

How we want to implement this:

1. When pod starts terminating, /drain_listeners?graceful&skip_exit
2. When application container shutsdown, SIGTERM Envoy.

The reason we need a skip_exit is due to the timing between the two. If we just do this without skip_exit if the application container takes > drain-time-s we will block new connections unexpectedly. While we could just raise the drain-time, this also impacts other draining (notably, LDS update), which may be unexpected.

Risk Level: Low
Testing: Added new integration test based on existing ones around draining
Docs Changes: Add docs to admin.rst