No metrics for docker containers



  • Hi
    I’m new to Netdata, the installation went smooth on a few machines, but there are no docker containers metrics.
    After Netdata Agent installation I see the containers on the resources list per machine (node), but none of them has data to show. I’ve even enabled docker metrics and set the collector plugin, tested debug mode okay, but still no data (https://learn.netdata.cloud/docs/agent/collectors/go.d.plugin/modules/docker_engine/)
    The host is Digital Ocean VPS with Ubuntu 18.04
    netdataproblem.jpg

    Any suggestions?!


  • Staff

    Hi. Those metrics are from cgroup.plugin (https://github.com/netdata/netdata/tree/master/collectors/cgroups.plugin#cgroupsplugin), not docker_engine.

    Netdata error.log is a good place to check

    grep -i cgroups error.log


  • Staff

    Thanks @ilyam8 for chiming in!

    Welcome to our community @TOMASZ KRÓL ! We will get to the bottom of this 💪



  • Thanks for your reply.
    I have examined error.log for cgroup but found nothing obvious to me.

    There are some rare errors like:

    2020-09-30 15:24:37: netdata ERROR : PLUGIN[cgroups] : PROCFILE: Cannot open file '/sys/fs/cgroup/blkio/docker/d551baaf687e1e4b737206d61bc205ebcbe4aac3ee185935ecf3949fd3fc529c/blkio.throttle.io_serviced' (errno 2, No such file or directory)
    2020-09-30 15:24:38: netdata INFO  : PLUGIN[cgroups] : CGROUP: unregistered network interface rename for 'vethac04743' as 'eth0' under 'traefik'
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : CGROUP: cgroup '/docker/1c02cdab29583ec7b85b99e41fe6329cfec5f211e3bbac62c654fd6bba52a677' has network interface 'veth9dec51a' as 'eth0'
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : CGROUP: registered network interface rename for 'veth9dec51a' as 'eth0' under 'traefik'
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.cpu' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.cpu_per_core' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.mem' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.writeback' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.mem_activity' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.pgfaults' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.mem_usage' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.mem_usage_limit' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.throttle_io' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:48: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.throttle_serviced_ops' on host 'docker-inet-one' already exists.
    2020-09-30 15:24:49: netdata INFO  : PLUGIN[cgroups] : RRDSET: chart name 'cgroup_traefik.cpu_limit' on host 'docker-inet-one' already exists.
    

    Any ideas?


  • Staff

    No ideas from this log. What about other charts? Do you see them updating?



  • Yeah, everything else works fine.
    Same happened on 3 other machines - I can see the docker containers on the list but without data.
    btw: should the docker containers metrics works out of the box?


  • Staff

    Yes, it should. There is time picker - do you access netdata via cloud? Did you try to open netdata dashbpoard directly?


Log in to reply