Proposed title of this feature request

Export OpenShift Monitoring Metrics via OTLP

What is the nature and description of the request?

As of today, [1] the only ways to extract in-cluster monitoring metrics are via PrometheusRemoteWrite or, get them via the /federate endpoint and transform them via an OTel collector.

While the first method (remote write) is very efficient, it has some caveats:

• Some backends or third parties don't accept remote write.
• Users cannot ship their application metrics directly in OTLP together with those coming from OpenShift directly, forcing them to maintain different gateways/exporters.
• Users cannot reuse some configurations from an existing OpenTelemtry collector between signals.
• The monitoring system might be optional very soon, which means that an alternative of recollecting these metrics without storing them in the OpenShift cluster will be needed.

The second method is not efficient and doesn't scale well.

The goal of this feature is to provide an option to export OTLP metrics via an OpenTelemetry collector (or a set) that is able to overcome the cited limitations.

Why does the customer need this? (List the business requirements)

To have an efficient and centralized way to integrate with third party backends/solutions that are OTLP native.

List any affected packages or components.

OpenTelemetry collector. Solutions might entail:

• Implementing a PrometheusRemoteWrite Receiver (OBSDA-854)
• Activating a PrometheusRemoteWrite Exporter in the Red Hat build of OpenTelemetry (OBSDA-481)
• Utilizing just the available Kubernetes receivers to provide a subset of OpenShift metrics, taking into account that the monitoring system might be optional very soon.

[1] https://www.redhat.com/en/blog/red-hat-openshift-opentelemetry-otlp-native-platform