Electrolysis/Release Criteria/Plugin Jank: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
(analysis is pending review)
 
(3 intermediate revisions by 2 users not shown)
Line 14: Line 14:
== BLOCKED_ON_PLUGIN_INSTANCE_DESTROY_MS ==
== BLOCKED_ON_PLUGIN_INSTANCE_DESTROY_MS ==


Analysis of these metrics from beta45ex2 is [https://github.com/vitillo/e10s_analyses/pull/25| pending review]
Analysis of these metrics from beta45ex2 has [https://github.com/vitillo/e10s_analyses/pull/25 been reviewed] and is now [https://github.com/vitillo/e10s_analyses/blob/master/beta45-withoutaddons/e10s_plugin_lag.ipynb published].
 
The results:
* Only three plugins are reported by at least 0.1% of experiment participants: Java, Silverlight, and Flash.
 
=== Java ===
Slight improvement on init lag, perhaps a slight regression on destroy lag.
 
=== Silverlight ===
Not enough responses to say whether or not the init lag's differences (perhaps a slight lag) are meaningful or just noise. Destroy lag shows an improvement on extreme ( > 2 seconds ) lag, and a regression on moderate (200-700ms) lag.
 
=== Flash ===
No appreciable improvement or regression on init or destroy.
 
== BLOCKED_ON_PLUGIN_MODULE_INIT_MS ==
 
This metric is pending analysis in {{bug|1261362}}.


= Bugs =
= Bugs =

Latest revision as of 14:44, 1 April 2016

Plugin jank is in the category "things that may regress up to 5% as a tradeoff for other benefits". Any regression in the *chrome* process should stop ship. Regressions in the content process are not nearly so important.

RASCI

  • Responsible: chutten
  • Accountable: bsmedberg
  • Supporting: aklotz
  • Consulted:
  • Informed: elan, cpeterson

Metrics

BLOCKED_ON_PLUGIN_INSTANCE_INIT_MS

BLOCKED_ON_PLUGIN_INSTANCE_DESTROY_MS

Analysis of these metrics from beta45ex2 has been reviewed and is now published.

The results:

  • Only three plugins are reported by at least 0.1% of experiment participants: Java, Silverlight, and Flash.

Java

Slight improvement on init lag, perhaps a slight regression on destroy lag.

Silverlight

Not enough responses to say whether or not the init lag's differences (perhaps a slight lag) are meaningful or just noise. Destroy lag shows an improvement on extreme ( > 2 seconds ) lag, and a regression on moderate (200-700ms) lag.

Flash

No appreciable improvement or regression on init or destroy.

BLOCKED_ON_PLUGIN_MODULE_INIT_MS

This metric is pending analysis in bug 1261362.

Bugs

No results.

0 Total; 0 Open (0%); 0 Resolved (0%); 0 Verified (0%);

Retrieved from "https://wiki.allizom.org/index.php?title=Electrolysis/Release_Criteria/Plugin_Jank&oldid=1125636"