1902229 - Use individual trace ressources for each trace type

Status: RESOLVED FIXED

(DevTools :: Debugger, enhancement)

Description

[Alexandre Poirot [:ochameau]]

Bug 1901643 allows to efficiently communicate bulks of ressources frow resource watchers.
We can leverage this so that the watcher actor can stop bundling sub ressource type with a custom throttling mecanism.

Doing this will help preserve the order of each of these trace events. They are currently fired via distinct arrays, which prevents knowing in what order they actually fired. The webconsole uses an additional sort logic against the timestamp. But this slow things down and that's something we can avoir doing in bug 1885054.

Comment 1

[Alexandre Poirot [:ochameau]]

Using independant resources highlights the cost of ResourceCommand architecture:
https://treeherder.mozilla.org/perfherder/comparesubtest?originalProject=try&newProject=try&newRevision=0127fecded12fb1ed7fe68b8cda98354672016e6&originalSignature=4763542&newSignature=4763542&framework=12&application=firefox&originalRevision=ddf1ba86d98006c55a137f00944fa3f80d37943b&page=1&showOnlyConfident=1
3% regression on client+backend and 15% on the UI

This will be mitigated by some fixes like bug 1902224 and bug 1902226.

Comment 2

[Alexandre Poirot [:ochameau]]

Attachment: Bug 1902229 - [devtools] Use a independent resource for each trace data type.

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/94f274c26fff
[devtools] Use a independent resource for each trace data type. r=perftest-reviewers,devtools-reviewers,nchevobbe,sparky

Comment 4

[Serban Stanca [:SerbanS]]

https://hg.mozilla.org/mozilla-central/rev/94f274c26fff

Comment 5

[Nicolas Chevobbe [:nchevobbe]]

Perfherder has detected a devtools performance change from push 94f274c26fff753919cd59c6d87f479e0ffdb165.

Regressions:

Ratio	Test	Platform	Options	Absolute values (old vs new)
22%	damp jstracer.ui-performance.DAMP	macosx1015-64-shippable-qr	e10s fission stylo webrender-sw	1,788.36 -> 2,173.53
20%	damp jstracer.ui-performance.DAMP	macosx1015-64-shippable-qr	e10s fission stylo webrender	1,784.90 -> 2,150.58
16%	damp jstracer.ui-performance.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender-sw	1,726.13 -> 1,996.94
15%	damp jstracer.ui-performance.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender	1,731.11 -> 1,987.40
13%	damp jstracer.ui-performance.DAMP	linux1804-64-shippable-qr	e10s fission stylo webrender	2,315.50 -> 2,606.56
11%	damp jstracer.ui-performance.DAMP	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	2,313.59 -> 2,564.81

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
15%	damp custom.jsdebugger.close.DAMP	macosx1015-64-shippable-qr	e10s fission stylo webrender-sw	16.66 -> 14.20
14%	damp custom.jsdebugger.close.DAMP	macosx1015-64-shippable-qr	e10s fission stylo webrender	16.58 -> 14.27
11%	damp custom.netmonitor.close.DAMP	macosx1015-64-shippable-qr	e10s fission stylo webrender-sw	17.25 -> 15.39
9%	damp custom.netmonitor.close.DAMP	macosx1015-64-shippable-qr	e10s fission stylo webrender	16.92 -> 15.34
8%	damp custom.jsdebugger.close.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender-sw	21.73 -> 20.04
...	...	...	...	...
2%	damp custom.netmonitor.open.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender-sw	173.17 -> 169.63

As author of one of the patches included in that push, we need your help to address this regression.
Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the patch(es) may be backed out in accordance with our regression policy.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

You can run these tests on try with ./mach try perf --alert 881

For more information on performance sheriffing please see our FAQ.