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
This aggregator dates from very early in the prototype. It was used to demonstrate a Summary-like metric aggregation, including the raw data used to calculate precise quantiles. However, it is a misleading type and produces confusion, particularly because the OTLP exporter turns the type into raw gauge points, meaning it exposes a route from Histogram instruments to Gauge data.
Twice today I answered questions from confused users who used the simple NewWithExactDistribution() aggregator selector and got a bunch of gauge points.
Proposed Solution
Remove this code.
Alternatives
Users can implement this aggregator easily. One of the "near misses" for this code is to support passing raw data through the OTel API directly to another metrics library. This aggregator doesn't support that use-case very well.
The text was updated successfully, but these errors were encountered:
Problem Statement
This aggregator dates from very early in the prototype. It was used to demonstrate a Summary-like metric aggregation, including the raw data used to calculate precise quantiles. However, it is a misleading type and produces confusion, particularly because the OTLP exporter turns the type into raw gauge points, meaning it exposes a route from Histogram instruments to Gauge data.
Twice today I answered questions from confused users who used the simple
NewWithExactDistribution()
aggregator selector and got a bunch of gauge points.Proposed Solution
Remove this code.
Alternatives
Users can implement this aggregator easily. One of the "near misses" for this code is to support passing raw data through the OTel API directly to another metrics library. This aggregator doesn't support that use-case very well.
The text was updated successfully, but these errors were encountered: