http_listener_v2 confuses prometheus clients by responding with HTTP status code 204

[yacoob]

Relevant telegraf.conf

[[inputs.http_listener_v2]]
  service_address = "127.0.0.1:9999"
  paths = ["/metrics/job/home.backup"]
  data_format = "prometheus"

Logs from Telegraf

n/a

System info

Telegraf 1.30.3 (git: HEAD@fd4af886)

Docker

No response

Steps to reproduce

1. Configure http listener expecting prometheus input (aka prometheus push gateway lookalike)
2. Push to it with a client using official golang prometheus client - in my case it's resticprofile

Expected behavior

Metrics are pushed without an error.

Actual behavior

With resticprofile, I get:

unexpected status code 204 while pushing to http://localhost:9999/metrics/job/home.backup:

Official golang client considers the push a success if the client receives status 200 or 202. Telegraf replies with 204, so the push is considered a failure (as in, an Error object is returned from Pusher.push(). FWIW, the metrics do go through. resticprofile also treats this as a warning, not a failure.

Additional info

No response

[powersj]

#15454