LoAF · Main-thread blocking
Long Animation Frames: time per visit the main thread was too busy to react.
LoAF verdict
v1.2.4 vs v1.2.3 · / · mobileNowv1.2.3
Improved
Main-thread blocking got 52ms shorter (−17.2%). A change only counts when it is at least 5% and at least 50ms. Here that means 50ms, so this one does.
shorter−50ms · too small to count · +50mslonger
Now · v1.2.4
251ms
Needs improvement
v1.2.3304msChange−52msGood up to200ms
Visits measured
191
on mobile, enough for a verdict
v1.2.3 · mobile212 / 30
v1.2.4 · mobile191 / 30
v1.2.4 · desktop69 / 30
Where visits landed
share of visits in each rangev1.2.344%54%
v1.2.459%40%
Good ≤ 200msNeeds improvementPoor > 600ms
More visits in the good range than on v1.2.3 (59% against 44%). About half of mobile visits block for 200ms or less.
Moved in the same release
Things that can keep the main thread busy. A place to look, not a cause.
Scripts behind long frames
The script that blocked longest on each visit with long frames · Mobile
| Script | Visits | Blocking p75 |
|---|---|---|
| www.googletagmanager.com/gtm.js | 92 · 30% | 260ms |
| Inline script | 43 · 14% | 180ms |
“New” means it never led on the previous release. Inline scripts are grouped together, and query strings are never collected.
LoAF across releases
Typical visit · each point is a deploy, ringed by its LoAF verdictGood ≤ 200ms
ImprovedRegressedUnchanged
LoAF by page
v1.2.4 vs v1.2.3 · mobile · pages need 30 visits on both sidesPageVisitsv1.2.3v1.2.4ChangeVerdict
191304ms251ms−17.2%Improved
137330ms273ms−17.2%Improved
164347ms287ms−17.2%Improved
55297ms246ms−17.2%Improved