OpenTelemetry metrics not conforming to semantic conventions

[jmezach]

While reading through this doc I noticed that New Relic seems to still use the "old" metric names when using OpenTelemetry. Specifically this list seems to use the old names according to this blog from the OpenTelemetry project.

This is also apparent in the product itself and not just in the docs. For example, on the Kubernetes tab of an OpenTelemetry service there are two charts called Service throughput (rpm) and Service response time (ms). Both of these charts base their data on the http.server.duration metric, but our services conform to the new semantic conventions so they report a metric called http.server.request.duration so these charts remain empty.

[github-actions[bot]]

Hi @jmezach 👋

Thank you for filing an issue! We'll triage your issue and let you know if we have questions, and then route it to the appropriate team so we can get it solved.

[rhetoric101]

@jmezach thank you for your helpful comments! We do need to update our docs to clarify that you can send in both the old and the new names. Also, we need to make some updates to our Kubernetes UI to accommodate this.

[ally-sassman]

Hi @jmezach , I appreciate your feedback, and I'll be fitting this into an upcoming sprint.

[ally-sassman]

Closing this issue because we removed metric references and replaced them with links to the OTEL docs (for example, the http.metrics reference on https://docs.newrelic.com/docs/opentelemetry/get-started/apm-monitoring/opentelemetry-apm-ui/#golden-signals