Björn Rabenstein

> I can sign the commits once the PR is approved, It seems signing will need a force push Yes, it changes the commits and therefore needs a force-push. However,...

This is very close to the top of my review queue now. Will have a look soon, hopefully today.

Looks good to me, but there are lint warnings now. (They are coming from a recently added check, but they are still relevant.) Also, DCO signing is missing. @shivanthzen could...

> Follow-up here would be to tackle the addition of exemplars! To clarify: This means making [withExemplarsMetric.Write](https://github.com/prometheus/client_golang/blob/291b0b0c426b6fb4025d8a3c2c8e7385026ba1d8/prometheus/metric.go#L180) aware of native histogram exemplars. Then, we only need to update the doc...

Maybe this could be bundled up with the idea of one binary for both prometheus and promtool. The "mode" of the binary could be selected with a command line flag...

Yeah, I would generally agree. But `prometheus --tool` is just much shorter to type than `prometheus --runtime.mode=tool`. Or in other words: This is a flag that people might actually type...

Played with it, and it seems to do what it should. (I haven't looked at the code.) Beyond the important restrictions already mentioned above (only works for `target_info`, not for...

> We should not use VectorSelector for the second argument in "info", it should be its own AST node. I'm torn here. On the one hand, I would like to...

Just quickly: The plan _is_ to make `info` work on all info metrics, and to automatically use the actual identifying metrics. The PR here is an MVP to test the...

WRT how to tell what the identifying labels are: One of the basic ideas is that this information is already present in the TSDB. Ideally, the targets exposing info metrics...