David Ashpole

I'm not sure if it would be prohibitively expensive or not.

One option would be for prometheus exporters to fail the scrape if the overflow attribute is set on any metrics. But unlike with the prometheus server's scrape limits, you wouldn't...

@jsuereth sorry for the confusion. I'm not suggesting we remove client-side limits, and I think those are useful. I meant to say that we could keep the current client-side limits...

@jmacd raised one potential concern: We want to add some OpenTracing fields (sampling priority?) to this specification, which have dots in them. This would potentially leave us with some fields...

@MrAlias I think that is about how exemplars are stored in the prometheus server. For clients: For histograms, Prometheus stores one exemplar per-bucket. When a new observation is made with...

This is still relevant.

@carlosalberto added

@open-telemetry/wg-prometheus

> target metrics would still be called target_info when ingested in prometheus? I.e. this isn't a breaking change, correct? Correct. This is keeping the same behavior, and is just trying...

@carlosalberto I would appreciate that. The only caveat is that the proto changes this is based on haven't been released yet, so this can't actually be implemented yet