It would be nice to have some sort of health endpoint that can be used to monitor the health of the stellar-core process. Right now, options for monitoring health include doing a TCP connect check to peer port or looking at the state field in the /info endpoint. Right now, when the main thread is blocked (e.g., during in-memory Soroban state rebuild) these checks don't work, which may cause spurious warnings.
It would be nice if the TCP connect check could be made to work because the TCP connect check allows health check to be run globally (e.g., through GCP). In comparison, the HTTP approach has the drawback of needing special access to the box (since the HTTP endpoints aren't designed to be externally robust).
It would be nice to have some sort of health endpoint that can be used to monitor the health of the stellar-core process. Right now, options for monitoring health include doing a TCP connect check to peer port or looking at the
statefield in the/infoendpoint. Right now, when the main thread is blocked (e.g., during in-memory Soroban state rebuild) these checks don't work, which may cause spurious warnings.It would be nice if the TCP connect check could be made to work because the TCP connect check allows health check to be run globally (e.g., through GCP). In comparison, the HTTP approach has the drawback of needing special access to the box (since the HTTP endpoints aren't designed to be externally robust).