Details

      Description

      From Slack Jie Yu:

      https://mesosphere.slack.com/archives/CCWLY6SLE/p1545260770151700

      i have a feature request. currently, we cannot view the disk usage of a Mesos persistent volume (only Mesos sandbox is shown)

      for instance, http://grafana-soak113s.testing.mesosphe.re:3000/d/mpivSaYiz/mesos-container-details?orgId=1&from=now-7d&to=now&var-service_name=monitoring&var-executor_name=prometheus&var-task_name=All&var-rate_interval=2m

      i wanted to view how much disk space Prometheus is using

      Mesos `/containers` endpoint does report disk usage for each persistent volume
      https://soak113s.testing.mesosphe.re/agent/60c42ab7-eb1a-4cec-b03d-ea06bff00c3f-S8/containers`

      disk_statistics` section contains disk limit/usage for each volume (including sandbox)

      if we can tag other volumes with persistence id that’ll be great

      e.g.: `disk_used_bytes{executor_name="prometheus",volume_persistence_id="xxxxxx"}`. For sandbox, there’s no `volume_persistence_id` (edited)

      Looks like this is even already a TODO for us  https://github.com/dcos/telegraf/blob/1.7.2-dcos/plugins/inputs/dcos_containers/dcos_containers.go#L239

      • Also add Perf, NetTrafficControlStatistics and NetSNMPStatistics
      • Add blkio cfq_recursive and throttling stats
      • Return all blkio metric ops, not just TOTAL

        Attachments

          Activity

            People

            • Assignee:
              gracedo Grace Do
              Reporter:
              gracedo Grace Do
              Team:
              Observability Team
              Watchers:
              Daniel Baker, Grace Do, Mergebot, Philip Norman
            • Watchers:
              4 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: