upstream/prometheus
1
0
Fork 0
mirror of https://github.com/prometheus/prometheus.git synced 2026-02-18 18:25:24 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions 7

fix(promql): native histogram custom buckets rate with empty buckets

Browse source 
While working migrating some dashboards I noticed that some
classic histogram and NHCB quantiles don't match up exactly.

Signed-off-by: György Krajcsovits <gyorgy.krajcsovits@grafana.com>
This commit is contained in:
György Krajcsovits 2025-11-27 15:21:27 +01:00
parent 7bb95d548c
commit bbd7acd1d3
No known key found for this signature in database
GPG key ID: 47A8F9CE80FD7C7F
1 changed files with 33 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

33
promql/promqltest/testdata/histograms.test vendored
View file

@ -708,3 +708,36 @@ eval instant at 0 histogram_fraction(-Inf, 1, series)
expect no_info
expect warn msg: PromQL warning: vector contains a mix of classic and native histograms for metric name "series"
# Should return no results.
clear
# Test what happens if NHCB has more buckets at the end of the `rate` interval
# than the beginning of the interval. In the test data, two new buckets show up
# after each other around midway through.
load_with_nhcb 15s
testhistogram_bucket{le="0.01"} 0x10 1x10
testhistogram_bucket{le="0.02"} 0x9 1 2x10
testhistogram_bucket{le="0.05"} 1x9 2 3x10
testhistogram_bucket{le="0.1"} 3x9 4 5x10
testhistogram_bucket{le="+Inf"} 3x9 4 5x10
testhistogram_count 3x9 4 5x10
testhistogram_sum 10x20
eval instant at 5m histogram_quantile(0.99, rate(testhistogram_bucket[5m]))
{} 0.043999999999999984
# Should result in the same value as the classic quantile.
eval instant at 5m histogram_quantile(0.99, rate(testhistogram[5m]))
{} 0.019799999999999998
eval instant at 5m rate(testhistogram_bucket[5m])
{le="0.01"} 0.0033333333333333335
{le="0.02"} 0.006666666666666667
{le="0.05"} 0.007017543859649122
{le="0.1"} 0.007017543859649122
{le="+Inf"} 0.007017543859649122
# Should have the same value projected from cumulative buckets to
# absolute buckets.
eval instant at 5m rate(testhistogram[5m])
{} {{schema:-53 count:0.007017543859649122 custom_values:[0.01 0.02 0.05 0.1] counter_reset_hint:gauge buckets:[0.003508771929824561 0.003508771929824561]}}