fix(om2.0): consistency, GaugeHistogram and special values#2854
fix(om2.0): consistency, GaugeHistogram and special values#2854
Conversation
There is no reset for GaugeHistograms , they represent a current and not a cumulative distribution. Reword the sentences copied from the cumulative histogram that are based on a different lifecycle and resets. Signed-off-by: György Krajcsovits <gyorgy.krajcsovits@grafana.com>
| Float and negative bucket values are allowed to make it possible to expose results of arithmetic operations on GaugeHistograms, such as the rate of change of a Histogram over time. | ||
|
|
||
| A GaugeHistogram SHOULD NOT include NaN measurements as including NaN in the Gsum will make the Gsum equal to NaN and mask the sum of the real measurements for the lifetime of the time series. If a GaugeHistogram includes NaN measurements, then NaN measurements MUST be counted in the Gcount and the Gsum MUST be NaN. | ||
| A GaugeHistogram SHOULD NOT include NaN measurements as including NaN in the Gsum will make the Gsum equal to NaN and mask the sum of the real measurements until NaNs are removed from the distribution. If a GaugeHistogram includes NaN measurements, then NaN measurements MUST be counted in the Gcount and the Gsum MUST be NaN. |
There was a problem hiding this comment.
What does it mean for something to be removed from a distribution? Is that something supported by client libraries today?
There was a problem hiding this comment.
I assumed it means exposer stops measuring NaN in any way e.g. recording rule is no longer broken etc.
What does it mean for something to be removed from a distribution? Is that something supported by client libraries today?
Not removing one observation, but reseting yes -- https://pkg.go.dev/github.com/prometheus/client_golang@v1.23.2/prometheus#MetricVec.Delete
There was a problem hiding this comment.
I think this would be enough:
A GaugeHistogram SHOULD NOT include NaN measurements. If a GaugeHistogram includes NaN measurements, then NaN measurements MUST be counted in the Gcount and the Gsum MUST be NaN.
There is no reset for GaugeHistograms , they represent a current and not a cumulative distribution. Reword the sentences copied from the cumulative histogram that are based on a different lifecycle and resets.