Kibana Metrics Logs Timeouts in some spaces

Hi folks,

I experience strange behavior in my kibana 7.6.2.

I have multiple different spaces, where I have not changed the default settings for metrics or logs dashboards.

In some spaces when I press on Metrics after long loading screen I see following:

In other spaces after a very quick loading I see the metrics as expected:

The settings for the metrics are exactly the same in both cases.

I am not even sure where to start digging for this

Anyone has an idea?

Thanks

Now after avoiding the timeout I manage to load the view (very long loading time) and the difference is now that all of the filebeats agents of the cluster appear as hosts too:

Filebeat events have also host.name. Looks like this is why it is considered as host in the metrics view (even though in the settings the pattern is defined like metricbeat-*) Why is that happening? Maybe it leads to very long loading time? How I can remove the filebeats from the host metrics overview?

Edit:

After I have changed the Index Pattern in the "Logs" settings to some index, which does not contain host.name the "Metrics" Pugin loads very quickly and shows all nodes as expected, but of course I am not able anymore to "dive" into the nodes logs for example.

Any Ideas how this can be solved?

:wave: What version of metricbeat and filebeat are you running?

@Kosodrom asked around the team and it would appear that in 7.6, we loaded filebeat nodes on the inventory screen by design. For performance reasons (as you've noted), we no longer do that. In the latest releases, we respect the index pattern defined on the settings tab and only load host for that index pattern.

2 Likes

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