perf(INP/desktop): deckgl layer-toggles header p75 11.7s — worst desktop target by two orders (#4537 Track D)

[koala73]

Problem

deckgl-layer-toggles > .toggle-header is the single worst desktop INP target by p75: 11,744ms (n=7) in the 2026-07-09 Sentry device split — two orders of magnitude past the next offender. Small n, but 7 near-12s interactions is a real defect, not noise. Surfaced during the #4537 refresh; currently unowned.

Where

The layer-toggles panel collapse/expand header — src/components/DeckGLMap.ts (and GlobeMap.ts mirror). Hypothesis to test FIRST: the expand/collapse triggers a synchronous deck.gl layer rebuild or a full updateLayers() commit in the click frame (cf. closed #4558 which chunked updateLayers with yieldToMain — this path may bypass that).

Diagnose first

node scripts/measure-dashboard-render-axis.mjs "https://www.worldmonitor.app/dashboard" --interact "selector:.deckgl-layer-toggles .toggle-header" --cpu-throttle 4 --json at DESKTOP viewport (--width 1365 --height 768) — worst-event-scoped phases + symbolicated top events name the culprit (prod keeps function names).

Acceptance

Sentry desktop webvital:inp bad-event rate for this selector → ~0; DebugBear desktop inpSelector row p75 < 500ms. Part of #4537 Track D.

[koala73]

Investigation update (no PR):

• Refreshed to current origin/main and traced both renderer handlers. The DeckGL collapse handler only toggles .toggle-list.collapsed, search visibility, and the chevron; the Globe mirror only toggles list/search display and the chevron. Neither handler calls render() or updateLayers().
• The requested clean-session Playwright capture could not reach the selector in this environment: after an initial map pointer interaction, the page remained .map-container.svg-mode rather than mounting DeckGL. This is consistent with the desktop renderer demand/capability gate, so a headless trace here cannot validate a DeckGL INP target.
• Existing DeckGL layer commits are already routed through the deferred heavy-data gate. The selector attribution alone therefore does not justify changing the collapse handler or introducing a broad render deferral.

Conclusion: the reported field outlier is still worth pursuing, but the current hypothesis is falsified and a patch would be speculative. Next evidence should be a production trace from a WebGL-capable desktop session with .map-container.deckgl-mode confirmed before triggering .deckgl-layer-toggles .toggle-header; use its worst-event phase and symbolicated stack to identify the concurrent render/layout source. No repository files were changed.