What did you do?
In the PromQL React UI, gave a PromQL expression where the Y values are over a small range, then view as a graph.
To replicate: e.g. foo < 0.10 > 0.096 or foo < 0.11 > 0.096
What did you expect to see?
Each Y-axis label should be distinct, with sufficient significant digits to distinguish them. e.g.
- 0.096
- 0.097
- 0.098
- 0.099
- 0.100
- 0.101
What did you see instead? Under which circumstances?
In the following graph, the data points range from 0.0965 to 0.0999:

I believe the Y axis label positions actually correspond to 0.0960, 0.0965, 0.0970 ... 0.1000, 0.1005, but all are showing as 0.10.
With values from 0.0965 to 0.11:

In this case I believe the Y axis positions correspond to 0.096, 0.098, 0.100, 0.102, ... 0.110, 0.112; but half of them show as 0.10 and half as 0.11
(Real-world use case which triggered this: alerts set at disks having less than 10% free space, then viewing the graph of the alerting expression when the values are only just dipping below and above 10%)
Workaround
Manually add * 100 (or other scale as required) to the expression you are graphing.

Reference
Possibly relates to old issue #3084
Environment
What did you do?
In the PromQL React UI, gave a PromQL expression where the Y values are over a small range, then view as a graph.
To replicate: e.g.
foo < 0.10 > 0.096orfoo < 0.11 > 0.096What did you expect to see?
Each Y-axis label should be distinct, with sufficient significant digits to distinguish them. e.g.
What did you see instead? Under which circumstances?
In the following graph, the data points range from 0.0965 to 0.0999:
I believe the Y axis label positions actually correspond to 0.0960, 0.0965, 0.0970 ... 0.1000, 0.1005, but all are showing as 0.10.
With values from 0.0965 to 0.11:
In this case I believe the Y axis positions correspond to 0.096, 0.098, 0.100, 0.102, ... 0.110, 0.112; but half of them show as 0.10 and half as 0.11
(Real-world use case which triggered this: alerts set at disks having less than 10% free space, then viewing the graph of the alerting expression when the values are only just dipping below and above 10%)
Workaround
Manually add
* 100(or other scale as required) to the expression you are graphing.Reference
Possibly relates to old issue #3084
Environment
System information:
Prometheus version:
Alertmanager version:
n/a
Prometheus configuration file:
n/a
Alertmanager configuration file:
n/a
Logs:
n/a