promql engine does not return expected results with mixed floats+histograms #14172
Closed as duplicate of#15177
Description
What did you do?
I am testing the behavior (with promqltest) of querying a series that has both floats and histograms.
For example:
load 1m
mixed_metric 1 2 3 {{schema:0 sum:5 count:4 buckets:[1 2 1]}} {{schema:0 sum:8 count:6 buckets:[1 4 1]}} 4 5 {{schema:0 sum:18 count:10 buckets:[3 4 3]}}
# I expect the following to work, but it Fails:
eval range from 0 to 8m step 1m mixed_metric
{__name__="mixed_metric"} 1 2 3 {{count:4 sum:5 buckets:[1 2 1]}} {{count:6 sum:8 buckets:[1 4 1]}} 4 5 {{schema:0 sum:18 count:10 buckets:[3 4 3]}} {{schema:0 sum:18 count:10 buckets:[3 4 3]}}
# The following does pass though.
# Where we expected Histograms, we got Floats from a lookback
eval range from 0 to 8m step 1m mixed_metric
{__name__="mixed_metric"} 1 2 3 3 3 4 5 {{schema:0 sum:18 count:10 buckets:[3 4 3]}} {{schema:0 sum:18 count:10 buckets:[3 4 3]}}
clear
# As another note, this only occurs when a float follows a histogram.
load 1m
mixed_metric 1 2 3 {{schema:0 sum:5 count:4 buckets:[1 2 1]}} {{schema:0 sum:8 count:6 buckets:[1 4 1]}}
# Passes
eval range from 0 to 5m step 1m mixed_metric
{__name__="mixed_metric"} 1 2 3 {{count:4 sum:5 buckets:[1 2 1]}} {{count:6 sum:8 buckets:[1 4 1]}} {{count:6 sum:8 buckets:[1 4 1]}}
# Fails
eval range from 0 to 5m step 1m mixed_metric
{__name__="mixed_metric"} 1 2 3 3 3 {{count:6 sum:8 buckets:[1 4 1]}}
What did you expect to see?
I expected to receive the Histogram point at 3m (in the first example).
What did you see instead? Under which circumstances?
Instead, I received a Float of 3. Because another float exists later on in the series, a lookback was used instead of the Histogram that exists at that point.
System information
No response
Prometheus version
No response
Prometheus configuration file
No response
Alertmanager version
No response
Alertmanager configuration file
No response
Logs
No response