PagerDeck for building services

The fault message arrives, just not at anybody

The plant did its job. It detected that the temperature is rising and it produced a message. What happens after that is unregulated in most buildings.

• A control system can send a mail or close a contact. At three in the morning neither reaches somebody who is asleep.
• One collective alarm for everything means either reacting to every fault, or eventually stopping to look at all.
• Whoever is on call is rarely allowed near the plant's configuration, and whoever is allowed near it is not on call.

Four chains from an on-call duty

The plant suggests an urgency. What becomes of it is decided by the chain in the recipient's account, by source, tags, text pattern and time of day.

• Cold room goes over its temperature limit

  Conditions (when)

  source: cooling-btags.all: cooling, threshold

  Actions (then)

  stop: true
  Critical

  Stock spoils in hours, not in days. This is the message somebody carries a duty phone for in the first place. Without Apple's approval PagerDeck delivers it as high, visible in the history.

• Door to a plant room is standing open

  Conditions (when)

  source: accessmatch.body: /door open/i07:00-18:00 mo-fr Europe/Berlin
  Silent

  During the day open is the normal state, because somebody is working in there. The message is recorded regardless. Outside the window the rule does not apply, and the plant's own suggestion stands.

• Standby generator running its test

  Conditions (when)

  source: generatormatch.title: /test run/i
  Silent

  The weekly test run is not a fault, but it looks exactly like one to the plant. A rule takes its sound away, and the history still shows that it happened.

• Lift reports a fault on its emergency call

  Conditions (when)

  source: lift-b1severity_at_least: error

  Actions (then)

  escalate_after: 15mthen_notify: device:bereitschaft-2
  High

  Breaks through a Focus, at any hour. If nobody acknowledges within a quarter of an hour, the message additionally goes to the second on-call device, which the rule names itself.

The conditions are written the way they stand in the rule editor. Evaluation runs top to bottom, and the first matching rule carrying stop ends the chain.

What usually sends

Anything that can send a mail or make an HTTP call. Every installation gets a source of its own, so that it can be ruled on and revoked on its own.

Frequently asked from building services

My control system can only send mail. Is that enough?

Yes. From Business, every source can be given its own address for the plant to send its fault messages to. Because such an address travels across other people's mail servers, it is treated like a key: revocable on its own, and with an allowlist that decides which sender may use it at all.

Does this replace a connection to an alarm receiving centre?

No, and that is not a cautious phrasing. A supervised connection is a tested transmission path that monitors its own availability. PagerDeck is a delivery to a phone and stands beside such a path, never in its place.

What happens if somebody learns a source's address?

Then they can inject messages for as long as the allowlist is empty. That is why it belongs filled in: once it is, PagerDeck only accepts mail whose sender is listed and whose domain passes the check. Everything else is refused during delivery, and each address can be withdrawn on its own without touching the others.

Does an alert get through when the phone is silenced?

The high level breaks through a Focus, but not through the silent switch. The critical level breaks through both, and for it Apple requires an entitlement granted per app plus the user's consent on the device. As long as both are not in place, critical falls back to high, visible in the history and without an alert being lost.