From 5639c8ed3c1d28d0d29fb4d94d5a80865074da50 Mon Sep 17 00:00:00 2001 From: Alexander Karan <47707063+AlexanderKaran@users.noreply.github.com> Date: Sun, 30 Aug 2026 16:12:01 +0800 Subject: [PATCH 1/2] Made Lighthouse Setup Clear --- packages/docs/src/content/docs/methodology.md | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/packages/docs/src/content/docs/methodology.md b/packages/docs/src/content/docs/methodology.md index cd17bce9..9102be39 100644 --- a/packages/docs/src/content/docs/methodology.md +++ b/packages/docs/src/content/docs/methodology.md @@ -160,6 +160,13 @@ implement the same small benchmark routes and data shape wherever possible, so the stats focus on browser rendering, server rendering, request-handler throughput, and load behavior for comparable production apps. +The client- and server-rendered browser tests use headless Chrome with +Lighthouse's desktop form factor, `throttlingMethod: provided`, and screen +emulation disabled. Lighthouse applies no simulated CPU or network throttling. +Requests use a local connection to a production server, so the results are +relative comparisons on the CI host rather than estimates for typical devices +or networks. + ### Framework Specific Notes - These runtime apps are not currently intended to measure static-site output. Astro's From 8ac565efe9640a048de00a662ab1565bef6c092e Mon Sep 17 00:00:00 2001 From: Alexander Karan <47707063+AlexanderKaran@users.noreply.github.com> Date: Sun, 30 Aug 2026 19:00:13 +0800 Subject: [PATCH 2/2] Tweak --- packages/docs/src/content/docs/methodology.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/packages/docs/src/content/docs/methodology.md b/packages/docs/src/content/docs/methodology.md index 9102be39..98e3f461 100644 --- a/packages/docs/src/content/docs/methodology.md +++ b/packages/docs/src/content/docs/methodology.md @@ -160,8 +160,8 @@ implement the same small benchmark routes and data shape wherever possible, so the stats focus on browser rendering, server rendering, request-handler throughput, and load behavior for comparable production apps. -The client- and server-rendered browser tests use headless Chrome with -Lighthouse's desktop form factor, `throttlingMethod: provided`, and screen +The client- and server-rendered browser tests benchmarks run directly on the Depot runner host and use the +host Chrome installation. They use headless Chrome with Lighthouse's desktop form factor, `throttlingMethod: provided`, and screen emulation disabled. Lighthouse applies no simulated CPU or network throttling. Requests use a local connection to a production server, so the results are relative comparisons on the CI host rather than estimates for typical devices