Auto-tools/Projects/JetPerf: Difference between revisions
< Auto-tools | Projects
Jump to navigation
Jump to search
(→Links) |
(→Links) |
||
Line 51: | Line 51: | ||
* https://wiki.mozilla.org/Jetpack/Release_Process | * https://wiki.mozilla.org/Jetpack/Release_Process | ||
* https://wiki.mozilla.org/ReleaseEngineering/How_To/Setup_Personal_Development_Master | * https://wiki.mozilla.org/ReleaseEngineering/How_To/Setup_Personal_Development_Master | ||
* http://hg.mozilla.org/build/tools/file/tip/buildfarm/utils/run_jetpack.py |
Revision as of 22:45, 10 January 2012
Jetpack Performance Tests
Running talos ts (startup) tests on Firefox for the addon builder set of addons
Overview
The Jetpack team maintains a set of APIs for add-ons to use and want to see how use of those APIs affect performance. They will provide a set of simple add-ons that use these APIs for testing. In the past we have brought up a set of unittests for jetpack on tinderbox. Recently there has been work to test startup performance of addons on the AMO website.
Tracking bug: https://bugzilla.mozilla.org/show_bug.cgi?id=717036
Bugs to be labeled with the [jetpack+talos] whiteboard entry
Goals
- Have a Tjetpack test which runs on checkins (and try server) to detect performance issues
- Track delta of each addon and make it graphable
Status Meetings
Jetpack weekly meetings Tuesdays @ 10:00am Pacific Time
Contacts
- Ateam: jmaher
- Jetpack: mossop
Plan of Action
- get list of addons from jetpack team
- get a list of commands to obtain the addons (checkout and build)
- create a Tjetpack test in talos
- proof of concept run to get results (by hand)
- determine graph server requirements (if any)
- review work with Jetpack team, ensure this meets their needs/goals
- integrate into talos
- staging and deploy talos.
Problems to resolve
- figure out how to integrate the download and compilation of jetpack addons
- this is in the jetpack unittests, but we have much less flexibility in our talos suite
- how to run all the permutations we care about (most likely 3 different tests):
- baseline + addon1
- baseline + addon2
- baseline + addon3
- how to display the results properly?
- is ts_addon1 data point good?
- is ts_addon2 ok?
- which branches should this run on?