so is your data a json file or influxdb? And what type of visualization do you want to use and which values do you want to plot?
Data is in json format
json file → telegraf agent → influxdata (db) → grafana
json has four fields - project name, jenkins job id, [lines, functions, branches], [hit, total, coverage] (last two fields are from gitlab repo)
parsed data from json file to influxdb are as below:
project_37_lines_hits,
project_37_lines_total,
project_37_lines_coverage,
Hey @yosiasz , I want table types of visualization of json data, colors can be ignored and want to plot all data jenkins job ids(37, 36, 35, 34, 33) has changed now to weekly time.
lines: hits, lines: total, lines: coverage
functions: hits, functions: total, functions: coverage
branches: hits, branches: total, lines: coverage
Rows:- project names should be rows headers and lines, functions, branches in sub header
Columns: - since(date) should be columns header and hits, total, coverage in sub headers
Getting back to the original question…
I’ve got a single query that returns a number of time series - 1, 2 or 4 depending on how the dashboard variables are set - system name and processor type being the determining variables. Each systems has two different processor types - CPU and GPU, for example.
I’ve set the panel to replicate by System name, so with a single processor type selected, I’m getting two panels - one for each system with a single time series shown on each one.
The problem comes when I select both processor types as well. The threshold/schema based color scheme I’m using for the time series means they end up the same colour in the legend, so I’d like to replicate the panel by processor type as well as by system name. There dosen’t seem to be a way to do this…
Given I have time series for SYS1/CPU, SYS1/GPU, SYS2/CPU, SYS2/GPU and both system name and Processor Type are in separate variables, I want 4 replicas of the graph - one for each combination of System Name and Processor Type to be displayed.
I guess I either want a second variable to replicate by, or an option to replicate by time series.
I am getting a similar issue after upgrading to grafana 12.2.0
My dashboard is a modification of well known ‘Node exporter Full’.
What I do not understand is that it was working perfectly well with grafana 7.3.7, and does not work at all after upgrading to 12.2.0.
My Dashboard declares several global variables this way:
job “node”(fixed value) Hidden
group label_values(node_uname_info{job=“$job”},group) Displayed
name label_values(node_uname_info{job=“$job”,group=“$group”},name) Displayed
node label_values(node_uname_info{job=“$job”,name=“$name”},instance) Hidden
machinetype label_values(node_uname_info{job=“$job”,name=“$name”},nodetype) Hidden
On dashboard displayed to user, only variables ‘group’ and ‘name’ are visible.
Selecting something for ‘group’ restricts the available values for ‘name’.
And finaly selecting something for ‘name’ determines the value of ALL other variables.
This worked exactlty as described with 7.3.7.
With 12.2.0, selecting something for ‘group’ has absolutely no impact on available values for ‘name’, and selecting something for ‘name’ does not change values of remaining variables.
I tried playing around with the different variables parameters:
- Regex
- Refresh=On dashboard load|On time change
- Selection options
But without success, I haven’t found a way to make this work as expected with 12.2.0 !
Why did you change behaviour of variables evaluation?
Is there any way to work around this issue?
Bug is known, and depends of your version of grafana, and underlying database.
Solution is described here : Prometheus: Label values variable query not returning correct set umbrella issue · Issue #91074 · grafana/grafana · GitHub