David Ashpole

/triage accepted /assign @rexagod

/assign @CatherineF-dev /triage accepted

/assign @dgrisonnet

> The primary "tool" we have to address this cardinality problem is the use of start_time to signal reset of an individual timeseries. Questions for the audience, considering cumulative exporters...

> What can they use to report a histogram, as it doesn't have an asynchronous version? They wouldn't be able to use the approach above today, and would need what...

@anuraaga I think I understand your confusion now. The OpenMetrics guidance to not expose timestamps is only meant to apply to metric timestamps, not exemplar timestamps. We can, and should...

From the OM spec: > in general, MetricPoint timestamps should not be exposed The section is only talking about timestamps on MetricPoints, not on exemplars

I've written a proposal related to this, which proposes not adding unit to the metric name when translating from OTel to prometheus: https://docs.google.com/document/d/16Wo-QHZLcKO0uFx97HPUJ-ZMFXSPviQ7Oc4-bShgcto/edit?usp=sharing Feedback is welcome. It would be especially...

From https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/compatibility/prometheus_and_openmetrics.md#metric-metadata-1: > Prometheus SDK exporters MUST NOT allow duplicate UNIT, HELP, or TYPE comments for the same metric name to be returned in a single scrape of the Prometheus...

/triage accepted /assign @dgrisonnet