Hello team,

Is there any reason why in kubernetes charts for latency (node/pods...) are empty? - the same case is visible in your demo.

How to get latency charts for kubernetes?



  • Hi,

    please send the logs:

    cd /home/lpar2rrd/lpar2rrd # or your LPAR2RRD working dir
    tar cvhf logs.tar logs tmp/*txt tmp/*json
    gzip -9 logs.tar

    Send us logs.tar.gz via

  • mgregorc_p4
    edited March 26

    OK, I will provide logs soon, but as I wrote - the same case is in your Demo - -> Containers Latency tab.

  • Any findings? logs were uploaded March 25.

  • we will provide you a fix this week.

  • OK, thx.

  • Any update?

  • Any update?

  • PeWu
    edited May 9


    we are collecting 'container_fs_io_time_seconds_total' from cadvisror. Description: "Cumulative count of seconds spent doing I/Os".

    We ran some IO workloads in containers and compared node sysstat data with cadvisor. It seems 'container_fs_io_time_seconds_total' is in nanoseconds rather than seconds.

    Fix is almost ready, we are testing it right now.

  • how is testing going?

  • IS the fix included in 7.90 release?

  • no, data provided by the kubernetes API is weird, or better, it is wrong as per us, it would have to be fixed on the kubernetes side

  • OK, but is there any action plan? current charts are unfortunately useless.

  • we know, but we do not know what to do, problem is not on our side, i.e. no any action plan

Sign In or Register to comment.