LearnNewsExamplesServices
Frontmatter
id14592
titleFM cockpit density evidence spike: real fleet-scale numbers before layout freezes
stateClosed
labels
enhancementdeveloper-experienceai
assigneesneo-opus-vega
createdAtJul 4, 2026, 4:02 AM
updatedAt2:43 AM
githubUrlhttps://github.com/neomjs/neo/issues/14592
authorneo-opus-vega
commentsCount3
parentIssue14560
subIssues[]
subIssuesCompleted0
subIssuesTotal0
contentTrust
projected
quarantined0
signals[]
blockedBy[]
blocking[]
closedAt2:43 AM

FM cockpit density evidence spike: real fleet-scale numbers before layout freezes

Closed Backlog/active-chunk-3 enhancementdeveloper-experienceai
neo-opus-vega
neo-opus-vega commented on Jul 4, 2026, 4:02 AM

Context

T0.2 gate leaf of Epic #14560 (map v1.1), carrying the #13444 graduation's density AC down to the cockpit surface: "the UX epic needs actual count/density evidence before freezing layout." Filed under the operator-escalated decomposition sprint (D#14561).

The Problem

The SSOT artifact mocks 6–7 agents, ~5 stream events visible, 2-column grid. Tonight's live reality already stresses those assumptions: 12 registered identities, message bursts of 40–90 per 25-minute window (observed 2026-07-04, wake-batch sizes), 15+ tickets filed per hour during a sprint. A layout frozen against the mock's density will break at real fleet scale — the #13444 review named this class explicitly (scale-to-a-glance needs numbers, not vibes).

The Architectural Reality

Evidence sources, all queryable today: the Memory Core roster (who_is_online — registered vs active vs benched counts), A2A volume (message timestamps per window from list_messages paging), lifecycle event rates (PR/review/comment timelines from the synced content), lane counts (D#14561 claims ledger + open epic assignees). The temporal-pyramid L2 records (#14433 orbit) will eventually make this a query; this spike hand-collects once.

The Fix

A short evidence doc (comment on #14560 or apps/agentos/design/ sibling artifact — placement per #14577's decision): agent-count distribution (registered/active/concurrent), events-per-minute percentiles for the activity stream (median + burst), lanes-per-agent, and the derived layout consequences (grid columns at N agents, stream buffer + collapse thresholds, health-bar categories). Each number carries its query/source (the #14430 falsifying-query discipline applied to design inputs).

Acceptance Criteria

  • Agent-count evidence: registered / active-in-window / peak-concurrent, with sources.
  • Activity-stream rate evidence: median + p95 events/min over ≥3 real windows (tonight's burst windows qualify), with sources.
  • Derived layout consequences stated for grid, stream, and health bar (the numbers the view leaves consume).
  • Every figure carries its falsifying query/source; no vibes-numbers.

Out of Scope

  • The views themselves (T2–T5 consume this).
  • Temporal-pyramid automation (#14433 orbit — this is a one-shot hand collection).

Related

Epic #14560 (parent) · #13444 (the density AC's origin) · #14577 (placement gate) · D#14561 (sprint).

Live latest-open sweep: latest 12 checked 2026-07-04T01:59Z (newest #14591); no equivalent. A2A in-flight sweep: clean for L1 leaves (lane-3 batch is Clio's, boundaries noted). Structure-map gate: N/A — evidence doc, no new ai/ placement.

Origin Session ID: a28f1415-37a4-4a5a-a3e4-b56adf6274ec Retrieval Hint: "FM cockpit density evidence agent counts stream rate layout freeze"

tobiu referenced in commit 7eb145a - "feat(agentos): FM cockpit control cluster — per-agent power toggle + whole-fleet + honest-state render (#14611) (#14892) on Jul 6, 2026, 6:56 PM