Intermittant Metricbeat data on K8s


(Tim Ward) #1

I've just started playing with Metricbeat on K8s, and am trying to make sense of the data that's coming in. To take the supplied "[Metricbeat Kubernetes] Overview" dashboard as an example, sometimes it shows all zeros rather than real numbers for the nodes, deployments, pods, etc, counts, and there are gaps in some (but not all) of the graphs.

All I did to the provided K8s configuration file was

  • take out the RBAC stuff as we're not using it
  • wind one of the data collection numbers down from 10s to 30s (maybe that was my mistake, but we're paying for the bandwidth?)
  • set up the Elasticsearch destination
  • add the dashboard setup

image


(Noémi Ványi) #2

By increasing the interval of data collection to 30s , it's possible that some samples do not make it in time, so a few buckets are empty. Could you try decreasing the sampling interval to 20s or 15s to see if the problem persists?


(system) #3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.