Hello, I am trying to deliver a Kibana dashboard providing counters' measurement data. There are two fields that identify counter measurement uniquely. DN(distinguished name) which describes counter measurement source(VNF/VNFC from which counter measurement has been scraped) and counter display name.Dashboard contains a control visualization, a date histogram visualization and a table (from search of kibana). The control visualizationhas 4 option list with parental relationship: VNF name, Distinguished name, Counter Group name,Counter display name. Aim is let user select through drop down lists counters and then provide a histogram with counters values in different time frames. Problem is that upon dashboard initialization and before user selects some filters through the drop down lists, histogram drawing causes max buckets exception since number of DNs and counters to be depicted are huge. Keep in mind we need combination of DN and counter name to identify counter uniquely since counter name could be same for different sources(DN). I am seeking a way to delay histogram painting until enough filters(DN,counter name) have been provided by user in order to narrow down amount of data to be painted in histogram. I tried to make two separate dashboards, one containing control visualization with filters and a second one containing histogram. i am struggling to find a way how i can pass two filters(DN,counter name) from first dashboard to second one. i just managed to add a link to counter name field(through index pattern URL functiolaity) redirecting to second "histogram dashboard" but i can pass only counter name parameter but not second one(DN). So if same counter name belongs to more than one DNS, all measurements are drawn and not of the specific DN. Is there any way to pass all filters provided by user to histogram visualization? Thanks in advance, Dimos
this request came up a few times in the past, there is an issue tracking it here: https://github.com/elastic/kibana/issues/26159
Unfortunately it's not possible to do this right now.
You mention the histogram vis is crashing because it tries to use too many buckets, can you reduce the "size" of the terms aggregations for the various fields to reduce the number of total buckets (maybe just showing top 5)?
It could be possible to show just counters from first 5 DNs in the list and this would solve the problem.But you can understand that probably user would be confused about logic behind drawing a portion of counters. Generally it could be helpful if feature delaying histogram painting until filters are provided is delivered at some time.
Thanks for your input.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.