What is the problem you are trying to solve?

Leftover from #11100 .
Related to #9072 .

Mimir should handle the created timestamp field in Remote Write 2.0 according to spec.

Which solution do you envision (roughly)?

If the value of the created timestamp field in RW2.0 is not zero, we should inject a zero sample (float or float/integer native histogram) at the created timestamp.

Out of order created timestamp is ignored without error, as well as created timestamp in the future.
Duplicate sample value is handled as with any other sample. Exact match ok, diff is an error.

Created timestamps were defined here: https://github.com/prometheus/proposals/blob/main/proposals/0029-created-timestamp.md

Remote Write 2.0 spec: https://prometheus.io/docs/specs/prw/remote_write_spec_2_0/

Handling created timestamps for native histograms: https://prometheus.io/docs/specs/native_histograms/#created-timestamp-handling

Have you considered any alternatives?

We already have an alternative with the silent zero approach introduced in grafana/mimir-prometheus#689.

However it is a custom solution that has some problems and isn't upstreamed.

Any additional context to share?

No response

How long do you think this would take to be developed?

Small (<= 1 month dev)

What are the documentation dependencies?

No response

Proposer?

No response