Luminous Prometheus exporter on Large clusters

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


Morning, List.
I recently upgraded one of my large (1k osd) clusters from Jewel to Luminous. The upgrade went wonderfully smoothly, and I've now got 3 mgrs running as well.

One of the things I was most looking forward to getting my hands on was the "native" prometheus exportor. I enabled it as per the documentation, and scraped it:

time curl sct-r1-r1-head1:9283 | wc -l
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 18.5M  100 18.5M    0     0   424k      0  0:00:44  0:00:44 --:--:--  297k

real    0m44.715s
user    0m0.224s
sys     0m0.562s

That's ... a *lot* of metrics. Just wondering how large of clusters this has been tested with. It seems to me I'm going to need to turn off some metrics and/or break them into different types of requests to get decent scrape times and small enough scrapes that I can provision a prometheus server to actually be able to ingest it. Has anyone else with a large cluster looked at this yet?

FWIW, the dashboard plugin performs fairly admirably as long as I don't go to the OSDs page...

This e-mail message and any attachments are only for the use of the intended recipient and may contain information that is privileged, confidential or exempt from disclosure under applicable law. If you are not the intended recipient, any disclosure, distribution or other use of this e-mail message or attachments is prohibited. If you have received this e-mail message in error, please delete and notify the sender immediately. Thank you.

Ceph-large mailing list

[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFS]

  Powered by Linux