// OPS WALL

NO TABLES OF ITS OWN.
EVERY APP’S PULSE.

The problem is not “a second Sigma.” The problem is a 60-second ops wall: five headline numbers (registered apps, active Kronos jobs, open SRM, NLG pool, token cost), a live grid that polls each blueprint’s /health, three charts, and a 20-row activity feed from SHL milestones and Herald. Nexus owns tables: []. Metrics prefer the Sigma API and fall back to Kronos/GHT/SRM/NLG/TKN tables. Login on /nexus/app is Flask-Login, no SWIP callback inside this blueprint. JSON uses mcp_or_login. No Nexus env vars. README still mentions Telegram helpers — tokens retired. This is read-only: you do not mute alerts or edit jobs here. That is Herald and Kronos.

Premier tables: [] 60s refresh Sigma-first
Open Nexus How it works
0
Owned tables
5
Headline stats
20
Activity rows
60s
Browser refresh
// HOW IT WORKS

Five steps that are all reads

If Sigma is down, fallback SQL still tries. If a /health 404s, the grid should show it — not hide it.

01
Login the shell
/nexus/app. Landing stays public and empty of live secrets.
02
Read the strip
Apps, Kronos jobs, SRM open, NLG pool, TKN cost.
03
Watch health dots
Polls registered /health endpoints.
04
Skim charts
App health doughnut, Kronos doughnut, SRM status bar.
05
Skim activity
20 events: SHL + Herald. Then go to the owning app to act.
Open Nexus
// STRIP

Five numbers, none invented — CSS mock

Zeros mean empty tables or Sigma fallback 0. Not an outage Lottie.

meltuc.tech/nexus/app — CSS mock, not a product screenshot
Appsregistered
Kronosactive jobs
SRMopen
// GRID

Health is HTTP, not a vibe

A red cell is a failed /health, not a designed alert.

meltuc.tech/nexus/app · health — CSS mock, not a product screenshot
Polleach blueprint /health
Period60s
Authlogin shell
// vs SIGMA

Wall vs registry

Sigma: 19 queries + threshold rules. Nexus: no tables, prefers Sigma API.

meltuc.tech/nexus/ vs /sigma/ — CSS mock, not a product screenshot
Nexustables []
Sigmasigma_alert_rules
Act elsewhereHerald/Kronos
// READ-ONLY

You cannot fix a job here

Click through. This page is a map, not a wrench.

// WHAT IT REFUSES

No writes. No SWIP callback. No public health grid.

Will not store a parallel metric warehouse. Landing does not poll 50 /healths for strangers.

// WHO THIS IS FOR

Admins who want a TV-friendly wall

Engineers with a MelTuc session. Not a statuspage.io for customers.

// FEATURES

Poll, prefer Sigma, don’t write

No invented incident commander.

📊
Stat strip
Five headlines.
🟩
Health grid
Per-app /health.
🍩
Charts
Health, Kronos, SRM.
📰
Activity 20
SHL + Herald.
🔌
Sigma first
Then direct SQL.
🚫
No tables
metadata tables [].
🔑
MCP or login
JSON dual auth.
⏱
60s refresh
Browser interval.
🛠
Read-only
Act in owning apps.
// GET STARTED

Open the wall. Fix things elsewhere.

Login for /nexus/app. This page does not live-poll the fleet for the public.

Open Nexus

Requires a MelTuc account. Create one free.

// SAID SLOWLY

What stays true over time

Platform Operations Center — real-time command dashboard aggregating health and activity. Health stays cheap. If this page disagrees with the signed-in app, trust the app — the product contract and tests enforce that rule.