MCP Healthchecks

[ggfevans]

Spurring off the mention of the Gateway deprecation, I'm wondering about the context of healthchecks for MCP servers.

I have a few MCP servers that can be somewhat flaky. This can cause some stability issues and token waste, particularly for Claude Desktop. One of the things that led me to this project was the need for some sort of health check dashboard to give me a canary in the coal mine that can tell me if a certain endpoint is flapping or just entirely dead. Obviously fixing the root issue is ideal, but we all know that uptime is never 100%

With the gateway soon to be deprecated, I'm wondering if that has bearing on health check functionality. I'm not sure if there's codebase overlap but the functionality of interacting with the MCP servers seems like it would overlap.

Anyways, I'm curious on your thoughts on healthchecks @tylergraydev and also wondering what other people are thinking or using here too.

[tylergraydev]

Nope. The health checks and ability to test the MCP calls yourself will remain untouched.