How do you keep an eye on a house or a warehouse you are not in?

The equipment that looks after an empty building is the equipment nobody looks at. It is in a cupboard, behind a door that is closed for the winter, in the corner of a warehouse where the light has been off since October. It works, until it does not, and the silence between those two states can last months.

What this is, and what it is not

It is not an alarm. An alarm detects an event and calls for help; that is a different job, done by different equipment, often with a contract behind it. Nothing described here replaces it.

What this does is quieter and, in its own way, more fundamental: it tells you that the things you are counting on are still there. That the connection is up. That the recorder still answers. That the storage box in the office has not been offline since the storm. It is the layer underneath the alarm — the one that notices when the alarm itself has stopped being able to call anybody.

What is worth watching in a building nobody is in?

• The connection. Everything else is downstream of it. From outside you learn that the building is unreachable; that alone answers most questions at seven in the morning.
• The recorder and the cameras, as devices that must answer. Whether they are recording is a separate question and a harder one — there is a whole guide about it — but a recorder that stopped answering on the 3rd of the month is a fact you want on the 3rd, not in March.
• The gate or door controller, which tends to be forgotten precisely because it works. A controller that has quietly gone offline keeps opening with the badges it already knows and stops applying anything you changed since.
• Network storage, which is where the recordings or the backups usually live, and which fails silently by design — most of these devices cannot raise an alert on their own, they can only be asked.
• The router and the switch, so you can tell “the building is offline” from “one device is dead”.
• Anything that publishes a number: a power supply unit reporting battery state, a device reporting its temperature. Where it exists, it turns guesswork into a reading.

If a device goes quiet, do you know why?

No: when a device stops answering in an empty building, the network tells you one thing only, that it stopped answering. It does not tell you whether the power went, whether a cable was pulled, whether the power supply died or whether the device crashed. Anyone who claims to distinguish those from the outside is guessing.

What you can do is reduce the guessing. If the whole building goes quiet at once, it is the line or the power. If one device goes quiet while its neighbours answer, it is that device. If a power supply unit publishes its state, you know which of the two it was. That is usually enough to decide whether you get in the car.

How does it work without opening anything?

It works outwards only: for anything reachable from outside, nothing has to be installed at all. For the devices inside — which is most of them — a small probe sits on the same network and asks the questions locally, opening only an outgoing connection, the way a browser does. Nothing is opened inwards: no port forwarding, no VPN, no fixed address. It runs on ordinary hardware, on a small single-board computer, or as an agent on a machine that is already there.

What to do this week

1. Walk the building, or picture it, and list what has an address. Most people find two or three devices they had forgotten entirely.
2. Watch the connection first, then each of those devices for simple reachability.
3. Decide what deserves a phone call — usually only “the building is unreachable” — and leave the rest as a message you read in the morning.
4. If something there publishes a number worth knowing, a temperature or a battery, watch that number too.
5. Write down what you expect. In six months that note is the only way to tell a real change from a vague memory.

You cannot be in two places at once. You can know, within minutes, when one of them has gone quiet.