I3140 — The mesh page's latency is measured against a recorded baseline before any target is claimed

✕BLOCKED wave 2 · p3 · verification profile · parallel safe

Part of prismatic-lumen — Prismatic Lumen: the owner oversees and moves the whole mesh from one page, which owns nothing.

Objective

Record, under rule project.performance-evidence and through a measuring command under scripts/ui and scripts/lib, warm status on a named route, cold and warm first paint, local event to screen, and subscription recovery for the mesh page, with hardware and percentiles; only then declare targets, as budgets of benchmark policy.

Why

The 2026-10-18 plan proposed latency targets for status, event to screen and recovery; the repository refuses 'fast' without a baseline, and ADR 0196 makes such targets budgets of benchmark policy. Critique of 2026-10-18 (latency-baseline-has-no-instrument): no instrument measures a /ui page's paint, event-to-screen or recovery. A baseline must come from its measuring command, and targets are benchmark-policy budgets. Proposed in ADR 0209 (Prismatic Lumen names the owner's mesh-wide view, and it owns no scheduler, event bus, store or API) (proposed; it carries no authority until the owner accepts it). Sources: chatgpt-majordomus:c296#12.

Current state

.ai/repo/benchmarks holds economics, observatory, pages and rust baselines; nothing for the mesh page.

Desired state

A baseline regenerated by its measuring command, and targets declared from it as budgets in .ai/repo/benchmarks/rust/policy.yaml.

Scope

• docs/PERFORMANCE.md
• scripts/ui
• .ai/repo/project/intents
• .ai/repo/benchmarks/rust/policy.yaml
• .ai/repo/benchmarks/rust/README.md
• .ai/repo/benchmarks/README.md
• .ai/repo/benchmarks/rust
• scripts/lib/ui-observatory.mjs
• scripts/lib/ui-routes.mjs
• scripts/lib/browser.mjs
• scripts/lib/ui.test.mjs
• scripts/lib

Out of scope

• Optimisation work
• A service worker or client cache

Dependencies

• I3138✕BLOCKED One page shows the whole mesh from the blackboard, peers and monitor with each figure's freshness, and recovers after a reconnect

Acceptance criteria

• A baseline with hardware, percentiles and routes is committed
• No target appears in prose before it
• The baseline is regenerated by a measuring command under scripts/ui and scripts/lib, beside scripts/lib/ui-observatory.mjs, with a declared route list; it is never edited by hand
• Warm status names the route it times; first paint is measured as a cold and a warm load, and no client cache is assumed
• I3140's why no longer quotes the proposed targets; they appear only as budgets in .ai/repo/benchmarks/rust/policy.yaml beside the UI budgets (ADR 0196, I2489), each citing the baseline it came from
• The criterion latency-is-measured-before-it-is-claimed is rebound in .ai/repo/project/intents/the-owner-oversees-the-mesh-from-one-page.yaml to evidence: test with the case that runs the measuring command

Validation

• majordomus plan validate

Evidence required

• mesh_page_baseline

Evidence

None recorded. Every token above needs a command or an artifact behind it before this issue can be completed; narrative is refused.

Timeline

started

—

verified

—

completed

—

Those three fields, the evidence above and the state of the dependencies are all the status is made of. There is no status field to disagree with them.

Canonical record: .ai/repo/project/issues/I3140.yaml. Read it back with majordomus plan show I3140.