The problem with watching 50 PBXs at once

An MSP running 50 customer 3CX servers has, in theory, all the monitoring it could want: health scores, MOS readings, trunk status, active-call counts, alert history. The problem isn't a shortage of data. It's attention.

A per-instance dashboard is exactly the right tool when you already know which instance you care about. It's exactly the wrong tool when you're trying to answer the only question that matters during a busy morning: is anything on fire right now, and if so, which one? Answering that with per-instance pages means clicking into 50 of them in sequence, or watching an alert inbox and hoping the important line doesn't scroll past while you're reading the one above it.

Alerts are essential — they page you when you're not looking. But a wall of text is the wrong shape for the thing a human eye is genuinely good at: noticing the one tile that's a different color from the other forty-nine. That's the gap a NOC wallboard fills. It trades depth for a single property nothing else gives you — everything, visible at once, readable from across the room.

This is the classic Network Operations Center pattern, and it earned its place. A NOC TV on the wall isn't there to replace your tooling; it's there so a tech walking past at 8:55am sees the red tile before the customer's office manager picks up the phone at 9:00.

What the 3CX NOC wallboard shows

Sikurd's NOC dashboard lives at /dashboard/noc and renders your entire fleet as a dense grid of tiles — one per 3CX instance, packed tight so a 50-server fleet fits comfortably on a single screen. Each tile is deliberately minimal, carrying only the signals you can read at a glance from ten feet away:

• A pulsing status dot. Green for online, red for offline, yellow for degraded, grey for unknown. The live ones pulse, so a healthy fleet visibly "breathes" and a dead tile sits conspicuously still.
• A large, color-coded MOS score — the focal point of the tile. Mean Opinion Score is the number that tells you whether calls actually sound good. It's rendered big and bold in the center of each tile, colored green / yellow / orange / red by band, because voice quality is the thing customers feel first and complain about loudest. A tile whose MOS has gone red is the call-quality canary, even while the status dot is still green.
• A health-score badge. The instance's overall composite health, at a glance, alongside the MOS.
• An active-calls count. How many calls are live on that PBX right now — useful context for whether a wobble matters this second.
• A critical-alert count. And if a tile has any critical alerts, the whole card gets a pulsing red border — the single loudest visual on the wall, so a genuine incident is impossible to miss even in peripheral vision.

That's the whole tile. No sparklines, no labels you have to walk up and read. Everything on it is sized and colored to survive being glanced at from across a room.

Worst-first triage sort

The detail that makes the wall actually work is the sort. Tiles are never ordered alphabetically or by customer name — that would scatter your problems randomly across a 50-tile grid and force you to scan all of it. Instead, Sikurd sorts strictly by how much each instance needs your attention:

1. Offline instances first. A dead PBX is the most urgent thing on the wall, so it's always at the top-left.
2. Degraded next. Up, but unhealthy — the things sliding toward trouble.
3. Then everything else by health score, lowest first. Among the instances that are nominally fine, the weakest float up so a developing problem surfaces before it becomes an outage.

The practical effect: the top-left corner of the wall is always your single biggest problem, and severity decreases as your eye moves down and right. You triage from the corner inward. On a calm day the whole grid is a wall of green pulsing dots and healthy MOS numbers, and the moment one instance drops, it climbs to the top and changes color — the wall reorganizes itself around your worst problem without anyone touching it.

Fullscreen NOC Mode

A wallboard is only as good as the display it lives on, so the NOC dashboard ships with a one-click "Enter NOC Mode" toggle. It triggers the browser's native fullscreen, stripping away the address bar, tabs, and app chrome until there's nothing on the glass but your fleet. The header shows the last-updated time and a quiet note that it auto-refreshes every 30 seconds; everything else is tiles.

Because NOC Mode opens in its own browser tab, you can dedicate a display to it and leave it there. Drop it on a wall-mounted TV in the ops area, or park it on the second monitor next to your PSA so it's always in the corner of your eye while you work tickets. There's nothing to babysit — the page re-fetches and re-sorts every 30 seconds on its own.

One deliberate non-feature is worth calling out: the NOC dashboard is a real-time tile wall and nothing else. There's no map, no scrolling alert feed, and no historical charts on this view. Those things matter, and Sikurd has them — but they live on their own pages, where you go when you're investigating. Stuffing them onto the wallboard would defeat its one job. The wall answers exactly one question — what is the current state of every PBX I manage? — and answers it from across the room.

How to run it on a wall display

Standing up a 3CX NOC TV takes about five minutes:

1. Pick a display. Any TV with an HDMI input and a small PC, a stick, or a Raspberry Pi behind it works fine. A spare monitor on a tech's desk works just as well for a one-person shop.
2. Open a browser to /dashboard/noc and sign in once. A read-only role is plenty for a screen that nobody clicks.
3. Click "Enter NOC Mode" to go fullscreen. The chrome disappears and the grid takes the whole panel.
4. Leave it. The 30-second auto-refresh keeps it current with no interaction. Disable the display's sleep timer so the wall stays lit through the day.

That's it. From then on the wall does the watching: a tech walking past sees offline-red or a collapsed MOS in their peripheral vision and acts on it before the ticket lands. For the deeper picture — why a tile went red, what the trunk is doing, the health trend over the last week — you click into the instance from a normal session. The wall's job is the glance; the rest of Sikurd is the investigation.

Where the NOC dashboard fits

The wallboard is the front door to a fleet, not the whole house. It pairs naturally with the rest of the platform:

• 3CX MOS scoring explained — what that big color-coded number on each tile actually means, and where to set your thresholds.
• Smart 3CX alerts with auto-resolve — what's behind the critical-alert count that lights up a tile's red border.
• Best tools for managing multiple 3CX servers — where a NOC wallboard sits in the broader multi-tenant toolkit.