Julien

That is super interesting. Can we see: - prometheus_target_scrape_pools_failed_total - prometheus_target_scrape_pool_sync_total Thanks!

I think the next step would be to get a go routine dump. https://prometheus/debug/pprof/goroutine?debug=2 (You can put the output in e.g. a [gist](https://gist.github.com)).

Do you have an idea for how long this instance is not updating its targets properly>

What comes into my mind is that we could get to this specific return: https://github.com/prometheus/prometheus/blob/44fcf876caad9a5d28b92a728d2c98c34015d377/discovery/kubernetes/endpoints.go#L194 Without sending an empty targetgroup for those endpoints. I'll need to dig firter. i do...

Indeed this can be done with `0.9*2*2^40`. My concern with introducing a specific function for bytes would be where to draw the line: should we also parse centimeters, milliliters, tons,...

I don't think we should do this. Prometheus only reads secrets from files. Username is not considered a secret. I am also concerned by the inability to read username and...

I will close this issue for now based on my previous reply.

Can you please provide an example?

gofumpt -w scrape/target.go -extra

Can you write a go benchmark?