Björn Rabenstein
Björn Rabenstein
@fpetkovski many thanks for picking this up again. Do you still have questions about how to proceed, or is it time for an in-depth review? (@bwplotka @roidelapluie @juliusv is one...
While in practice most counters on the same target reset at the same time, that's not generally true. Following cases come to mind: - Native histograms reset on purpose now...
Hasn't happened in a while. Will close this (although a proper solution would be better of course).
"Scaling" histograms (via multiplication/division by a scalar) sounds like a reasonable thing to do. However, I vaguely remember that I hit some roadblocks when I wanted to implement it a...
This has been implemented a while ago (including negative histograms).
Even if we won't fully support those two new features, we at least have to handle them gracefully. Currently, we just error out with `unknown metric type GAUGE_HISTOGRAM` in the...
The clearest way to not use the collectors already pre-registered with the default registry is indeed to create your own registry (which is anyway recommended for any non-trivial Go program...
I'm not sure if there is anything to gain here. The issue you are talking about was deliberately breaking, so a CI workflow would only have told us what we...
You can always use build tags to gate features that only work with more recent Go versions. Supporting the last 3 Go versions used to be very little overhead. There...
Grafana Agest apparently has its own config setting for this. Why do you then need a flag for Prometheus proper?