Due to bad mock implementation the TestRemoteWriteHandler_V2Message cases for OOO or duplicate sample within the same remote write message was resulting in bad request partial error (see the tests before removal of those test cases in #17838).

This is however not true. TSDB storage does not test anything around ordering for a single Appender batch (too expensive), so test was not showing reality, thus removed.

This issue is to:

• investigate if we need to add such extra validation e.g. before appending
• learn what happens after we commit such OOO data
• update compliance/proto accordingly if needed https://github.com/prometheus/compliance/tree/main/remotewrite