You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is a stand along feature but could serve as an entry point to --interactive mode in Perf-X #1143.
The basic idea is figuring out which types of resources have high performance impact so we can give some advice to improve site performance.
We can use some heuristics to find out what type of resources could potentially has high performance impact and run some tests blocking those resources then make suggestions based on the results.
Currently, this is still in experiment phase. I'm trying to figure out which types of resources are likely to have high performance impact and low rendering / functionality impact (so it would be meaningful for us to run the tests without users input).
Update: I did some research and decided not to implement Smart Insights mainly due to its potential low accuracy and the time limit of my internship. Here is the link to the doc: https://docs.google.com/document/d/1BO73_Bk6qr9XuJCBlDollg1SW39t_i5AEnlHI3eQofQ/edit?usp=sharing
The text was updated successfully, but these errors were encountered:
weiwei-lin
changed the title
Smart Insight - Run performance experiment automatically in the background.
Smart Insight - Run performance experiments and suggest changes.
Jan 24, 2017
weiwei-lin
changed the title
Smart Insight - Run performance experiments and suggest changes.
Smart Insights - Run performance experiments and suggest changes.
Feb 7, 2017
This is a stand along feature but could serve as an entry point to
--interactive
mode in Perf-X #1143.The basic idea is figuring out which types of resources have high performance impact so we can give some advice to improve site performance.
We can use some heuristics to find out what type of resources could potentially has high performance impact and run some tests blocking those resources then make suggestions based on the results.
Currently, this is still in experiment phase. I'm trying to figure out which types of resources are likely to have high performance impact and low rendering / functionality impact (so it would be meaningful for us to run the tests without users input).
Update: I did some research and decided not to implement Smart Insights mainly due to its potential low accuracy and the time limit of my internship. Here is the link to the doc: https://docs.google.com/document/d/1BO73_Bk6qr9XuJCBlDollg1SW39t_i5AEnlHI3eQofQ/edit?usp=sharing
The text was updated successfully, but these errors were encountered: