if i have a Master-Dashboard with a lot of datatables and i filter the data step by step.
Finally i get a Line Chart with the clicked filters.
With this workaround, you don't have to write the filter by hand on the "Visualization"-Site
Can i save the filtered Chart and implement this chart in another Dashboard?
Can i save the filtered Chart and implement this chart in another Dashboard?
It's a bit clunky, but yes, you can. As you are using the tables to filter data, it is creating filters in Kibana, near the top, under the search bar. You can then pin those filters, which will cause them to follow you around to Discover and Visualize. If you then edit the visualization on the Dashboard, or otherwise open it in Visualize, you can save it as a new Visualization (or, if you'd like, overwrite the existing visualization) and all those active filters will be saved as part of that new visualization.
Do you also know any possibillity to trigger the "dynamic" chart on the dashboard, which you modify with adding the filter with the tables, so that it will only monitor the values a special filter was added.
Now the Problem is, when i have a master dashboard with a lot of tables, where i can add step by step a filter, the dynamic dashboard makes first an average over all documents in this time range.
I'm not sure I understand what you're asking to do, but I think you're asking if you can see the records that are being used to generate a visualization. So that, as you apply filters and narrow things down, you can also see the records in Elasticsearch that match those filters.
If that's right, the answer is mostly no, but it kind of depends what you want to see. If you have a saved search (from Discover) in your dashboard, it'll show the matching records.
However, if you have a visualization (Table, metrics, bar charts, etc), then no. The reason being that those things are created using the aggregate data. You mentioned average, but it could be count, sum, media, percentile, etc. It's a search, using aggregations, so the data coming out of Elasticsearch doesn't include any of the underlying records, but the aggregated results of the search on your data.
Let me know if I misunderstood what you are asking for. I'm not sure what you mean by "dynamic chart."
I want to create a Master-Dashboard with huge usability. My data i want to monitor needs 4 filters, so that i can monitor the value, i want to see.
I created a kind of a treeview with 4 tables and now the user can add the needed filters click by click. Another part of the dashboard is a line chart, which monitors the average of all documents that are selected.
Now the user clicks the "treeview" till the wanted value and the line chart refreshes an monitors the values over the time.
And my problem is, if nothing on the treeview is selected, however the line chart does monitor a nonsense line (the whole datapool) and this costs a lot of performance.
Can you deactivate the chart till the last filter is set on the "treeview"?
I can't think of a way to stop that line chart from doing the initial rendering of all of the data. At least not in a way that makes a lot of sense. I can really only think of two possible solutions, but they're both bad.
You could save the dashboard in a default state with 1 filter applied, the most common filter (assuming you have one), so that at least for some use cases that initial render wouldn't be a wasted query. But then users would have to select the other filter they actually do want, and unselect the default one, so it's not great.
The only other thing I can think of would be worse. Basically, apply a filter that matches nothing, and have it enabled by default. This way the initial render will have 0 results, and nothing would be rendered. That doesn't really solve the load on the Elasticsearch side though, as the query would still happen. Also, when the dashboard initially loads, it'll look empty.
Is there a reason you don't just create 4 different dashboards, 1 for each filter, and simply start with the dashboard view that you want?
Thanks for your advices. The problem is that I have thousands of special values I can monitor. So the filter choice makes no sense.
The 4 tables are for navigating to different layer. I have a lot of devices with a lot of sensors. And I navigate with the tables to the bottom layer, the value of the sensor. The line chart is a kind of checking that there are values. And finally I can save this chart as special chart for this device, this sensor and the special value.
Mix of comfortable chart creation an control if there is data on this sensor
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.