I have the stock Metricbeat System dashboard for a host running very well but would like to add a logstash based visualization that presents java heap memory stats. I can see that the visualizations are based on the query for the specific beat.name value and I assume the whole dashboard looks to that data source and so the added visualization using the logstash index never gets populated.
So I have seen a couple of requests for things like nested dashboards, or the ability to change dashboard indices but are there any practical solutions to "mix and match" indices on a dashboard at the moment?
I guess I could create a java only dashboard and hyperlink to that from the main host dashboard at a push, or would grokking in a field named beat.name with the relevant hostname value work?
You can definitely create a dashboard and add whichever visualizations you want for that particular dashboard which can belong to a single index pattern or multiple indexpatterns.
It does come with some limitations. We cannot choose a indexpattern per visualization in dashboard yet. But that is coming soon.
Many thanks and yes it does, looks like I can leave the sample dashboards in place and then create a new dashboard dedicated to the various visualizations that can contain mixed indexpatterns with a hyperlink to join them, going forward the indexpattern per visualization would be the answer but this is the answer for now,
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.