You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
In regards to the issue #6376, the metric builder does not assign staleNaN values to the histogram and summary values for failed scrapes from Prometheus scrape loop. Therefore, the proposed possible solution is to use the OTLP format in the OTLP Prometheus Receiver with datapoint flags as a staleness marker for metrics with no-value present to address this issue.
Accordingly, if the metrics in the Prometheus Remote Write Exporter contain the OTLP flag that indicates MetricDataPointFlagNoRecordedValue, the Prometheus Remote Write Exporter should be able to handle and respond to the OTLP flag for staleness. This allows end-to-end compatibility in the collector pipeline with the use of OTLP native staleness marker flag, from the Prometheus Receiver to the Prometheus Remote Write Exporter.
Describe the solution you'd like
A possible solution is to check every data point for all metrics data types for the OTLP MetricDataPointFlagNoRecordedValue flag, and set the sample value of the data point containing this flag to the StaleNaN value. This ensures the Prometheus Remote Write Exporter to handle and respond to the OTLP flag for staleness in the collector pipeline.
Is your feature request related to a problem? Please describe.
In regards to the issue #6376, the metric builder does not assign
staleNaN
values to the histogram and summary values for failed scrapes from Prometheus scrape loop. Therefore, the proposed possible solution is to use the OTLP format in the OTLP Prometheus Receiver with datapoint flags as a staleness marker for metrics with no-value present to address this issue.Accordingly, if the metrics in the Prometheus Remote Write Exporter contain the OTLP flag that indicates MetricDataPointFlagNoRecordedValue, the Prometheus Remote Write Exporter should be able to handle and respond to the OTLP flag for staleness. This allows end-to-end compatibility in the collector pipeline with the use of OTLP native staleness marker flag, from the Prometheus Receiver to the Prometheus Remote Write Exporter.
Describe the solution you'd like
A possible solution is to check every data point for all metrics data types for the OTLP MetricDataPointFlagNoRecordedValue flag, and set the sample value of the data point containing this flag to the
StaleNaN
value. This ensures the Prometheus Remote Write Exporter to handle and respond to the OTLP flag for staleness in the collector pipeline.Additional context
Related issue: #6376
cc @alolita
The text was updated successfully, but these errors were encountered: