Open matschaffer opened 8 years ago
No I'm not sure how I would do observations with mtail directly. I assume you're interested in this for something like Prometheus where it's lossy between scrapes.
Exactly. And I'm also not sure :) But prometheus clients have histogram buckets so seems like there must be some approach to it. Will let you know if I end up digging further.
That's great -- I'd appreciate knowing if you come up with something!
Thanks for starting this. Looks really useful in helping to understand our insight into haproxy.
Have you been able to use this to get something like a 99th percentile for requests going through haproxy?