Indicator latency using le shows wrong availability and error budget

[sebastiangaiser]

Hey, I got a weird error in the Pyrra UI for SLOs using the latency indicator containing a le in it like here.

When using le=1.0 I get the following in the UI:

When using le=1 I get the following in the UI:

When using PromQL, I only get metrics containing e.g. le=0.1, le=0.2, le=0.3, le=0.5, le=1.0, le=2.0, ...