Bartlomiej Plotka

Thanks for more info. Was wrong about https://github.com/prometheus/prometheus/pull/17441 - it wasn't yet included in 3.8. The only commit that changes something that touches the RW1 sending path since 3.7: https://github.com/prometheus/prometheus/commit/f070e3535898408901d84f5c9aef386cd5f4ac15c...

Thanks everyone for amazing data. Kudos to @kgeckhart who did the main investigation and found (and [fixed](https://github.com/prometheus/client_golang/commit/7ba246a648ca4e294ca008d95b6fcc8df2f9c255)) the root cause. 3.8 fix (and summary of the root cause) is [here](https://github.com/prometheus/prometheus/pull/17683),...

Fixed in https://github.com/prometheus/prometheus/releases/tag/v3.8.1

Hi! Thanks for innovating here 💪🏽 I presume this is about using generics for label values type safety -- in the relation to defined label names. > Currently having unsafe...

Here is what I meant as an alternative for type safety (one does not exclude another one) https://github.com/bwplotka/metric-rename-demo/pull/1

Yup, that's a known problem for the `Const*` metric users. A great showcase is in [cadvisor](https://github.com/google/cadvisor/blob/be773c0a2b04536a818f79339bf89c82a74abb02/metrics/prometheus.go#L1879) also impacted by this. Kube state metrics get rid of client_golang for this reason...

Another solution is along our ideas for generated flow https://github.com/bwplotka/metric-rename-demo/pull/1 - we could create APIs that allow providing LabelPairs directly, so users can keep them ready to use.

> How could those headers help with partial write handling if it is unclear from them which samples were written? If I am sending series A, B, C, and B...

Ack, I agree, [so no change](https://github.com/prometheus/prometheus/issues/16945) and I mark this as done, thanks!

I added another point to protocol, but it could be tackled in 2.1 or so. If we would like to implement the delta type metric further, we need a solid...