Julien

A quick glance at the link shows me that we could accept it if there are enough tests and if we do not take configuration and secrets from environment variables....

What about we add an alert_relabel_configs to alertmanagers ? In addition of a global one in alerting

The AWS code is in common because it is also used in alertmanager. This code won't be used in alertmanager to the best of my knowledge

Prometheus should start indeed in this case . We should not make API calls while reloading the config Le mar. 16 mai 2023, 16:51, Chris Marchbanks ***@***.***> a écrit :...

We could add a /targets-metrics endpoint that would expose the targets with original target labels as individual metrics with a value of 1 (or 0 if the target is "dropped")....

We could have an http option expose_meta_labels ?

> So as discussed in person at the Kubecon, do we put this feature under [this](https://prometheus.io/docs/prometheus/latest/feature_flags/#extra-scrape-metrics) feature flag ? This can also be used to convey **scrape_interval** label We have...

Could you please try with a more recent Prometheus release?

Does using the env variable GODEBUG=netdns=cgo Help here?

It sounds like the issue you were experiencing has been resolved in Prometheus versions 2.40.7 and 2.41.0. If you are not already using one of these versions, upgrading to the...