Skip to content
Switch branches/tags
This branch is up to date with master.

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

RAGE, Results And Graphing Engine


A new framework being developed by the performance team to analyse the data collected, and display it as dynamic graphs on a webpage.


Test runs (jobs) are run on specific machines using specific builds of the product. Each test job can contain multiple test cases, each of which can contain multiple scales of measure (SOMs). Each such SOM can product multiple measurements.

Since the framework is fairly generic, its database schema is not trivial:

UML-like database schema

  • tiny_urls: A simple table for tiny URL support.
  • test_cases: Test case fully-qualified names and descriptions.
  • soms: SOM (scale of measure) definitions, where each SOM is part of a specific test case.
  • builds: Product build definitions. Each build is associated with a (source-control) branch, build number, and potentially a custom build tag.
  • branch_order: Ordering among different branches. RAGE is currently configured to ignore this table, ordering branches alphabetically.
  • jobs: A list of test runs (jobs). Each job is associated with a product build, and can optionally remember the original test command.
  • tc_config_<tc_fqn>: A table for each test case type. This table enumerates the relevant configuration option permutations for a specific test case.
  • som_config_<som_id>: An optional table for each SOM in case a SOM is associated with configuration options not captured by its test case.
  • machines: Test machine definitions.
  • tc_config: A table that provides non-build metadata for test case runs.
  • measurements: The table of measurements. Each measurements is associated with a test job, test case configuration, SOM, and (optionally) SOM configuration. The field result_id is used to allow multiple measurements for each unique permutation of the stated fields.
  • standard_builds: Friendly names for specific product builds.

We also provide the SQL to generate the main tables: [blob], [raw].

Tools used

Libraries used

Dealing with Git submodules

After cloning this repository, you need to execute the following command in order to fetch all Git submodules (e.g. ocaml-sql):

git submodule update --init

To obtain the latest changes for all submodules, use:

git submodule foreach git pull

To make changes in a submodule:

  • cd to the submodule, and make, commit, and push changes;
  • cd to the supermodule, and commit and push (the changed reference to submodule).

Development instructions

The program can be built on a system where OCaml, OMake, Core, and postgresql-ocaml are installed.

To run the program (omake run), a postgresql database must be running. The settings for accessing the database can be changed in OMakefile.

Installation of the program as a web service (omake install) simply copies the generated distribution (omake distro) into a web-server documents directory (/var/www by default). The installation currently assumes that flot files can be found in a sub-directory of the install directory.

Any errors will be shown in the web-server's error log (e.g. /var/log/apache2/error.log).

Citrix-internal instructions

A Citrix-internal machine that satisfies the above conditions is called perf. Make sure that your ~/.bashrc includes the following lines:

export PATH=/usr/local/perf/js/bin:$PATH
export MANPATH=/usr/local/perf/js/man:$MANPATH

This makes sure that you are using the correct versions of OCaml tools.


Results And Graphing Engine




No releases published


No packages published