Exposing Wasm VM state in a health check endpoint

*Title*: *Exposing Wasm VM state in a health check endpoint*

*Description*:
Any unrecoverable error in the Wasm VM will panic it, and all coming requests will get 503.
Users have to reload the wasm filter or restart the envoy process to eliminate the issue.

I didn't find any specific mechanism/API from Envoy documents that can help users identify the status of the Wasm VM.
This is useful for building a liveness probe to have the envoy process restart without any human interaction.

Thank you!

No files or tests are named in the issue. Start by locating Envoy's Wasm VM handling and health-check endpoint entry points, then determine how the endpoint can expose an unrecoverable VM state for a liveness probe; done means users can detect the failed VM and trigger an Envoy restart or filter reload.

Written by the indexing model from the issue text.