The infrastructure I'm deploying into employs a prometheus metric scraping proxy that connects to prometheus endpoints using a socket level client instead of an HTTP client.

My micronaut/jetty services are fine when deploying into this infrastructure, but micronaut/netty services don't get metrics picked up. Digging into the problem a bit, if I try to connect the the prometheus endpoint of a micronaut/netty service when it's running locally or deployed with a socket level client (ex. nc), I get HTTP 401. But if I repeat the test using an HTTP client (ex. curl) passing the same header data, I get metrics as expected.

Has anyone seen this or similar behavior from micronaut/netty HTTP endpoints and most importantly know of a workaround? I don't like my chances of convincing my infrastructure vendor to rewrite their scraping proxy to use an HTTP client in a reasonable amount of time.