Fuego wiki

Login or create account

chart config.json in 'raw' format

The file chart_config.json is defined for each test.  This file controls
how charts are drawn, in the Jenkins interface for the test it is associated

= schema =
chart_config.json holds a single object, with attribute pairs describing
values for the configuration of the charts for a test.  The following
pairs are supported:
 * '''chart_type''' - this indicates the type of chart to present in the Jenkins interface for a test job
   * It's value must be one of:
     * '''measure_plot'''
     * '''testcase_table'''
     * '''testset_summary_table'''
     * '''measure_table'''
 * '''match_board''' - if true, this indicates that a job page in Jenkins should only display the results for boards that match the job.
   * By default, Fuego displays the results for all boards on a job page
for a test.  This is to allow users to compare results between boards.  However, it is often desirable to only show the data for a single board (the one that the job actually refers to).
   * possible values are:
     * '''true'''
     * '''false'''

 * '''measures''' - this is a list of measures for which plots will be drawn
   * Each measure is specified by it's tguid, which must be specified in full (see [[Fuego naming rules]])
   * If no measures are listed in the chart_config.json file,
all the measures produced by the test will be plotted.

The purpose of the "measures" field is to limit the charting to only a few
important, or characteristic, measures.  Some Benchmark tests save many 
results, and the user may want to focus on only a few measures that they
are specifically interested in.

Here is an example, from the test Benchmark.cyclictest:
    "chart_type": "measure_plot",
    "measures": ["default.latencies.max_latency",

= Defaults =
If a test has no chart_config.json, then default values are used, as follows:
 * For Benchmark tests, create a measure_plot for all measures found in any run of that test.
   * Each measure_plot has one measure, shown with values for all boards that have had that test run.
   * The value and reference value (threshold) for each measure are plotted relative to each run of the test (by Jenkins build number).
 * For Functional tests, create a testcase table for each board.
   * Each testcase table has the result status for each result, and a set of summary lines at the bottom of the table, for each run of the test (by Jenkins build number)
 * For Functional.LTP, create a testset_summary_table for each board
   * Summary counts of pass/fail/error/skip status are shown for each testset (collection of test cases) for each run.

= Planned features =
Additional features are planned for future releases of Fuego, including the
 * additional chart_types:
   * testset_summary_plots - a plot of summary data by test set
 * control over measure grouping:
   * the ability to place multiple measures in the same plot
 * control over board grouping:
   * the ability to only show a single board, or specific groups of boards, in single plots
 * control over header data
   * the ability to customize the meta-data placed in table headers

= See also =
 * See [[Jenkins Visualization]] for more information about the charts that are configured by this file.

TBWiki engine 1.8.3 by Tim Bird