- Joined
- Jul 16, 2009
- Messages
- 9,471
- Reaction score
- 15,843
I have mentioned this to @Gaël and @Matthias Gross before.
The problem is well known in automation and monitorin: when a sensor hovers around a setpoint, you get endless nuisance alarms. Hysteresis doesn’t solve the problem. Other industries (fire, security, process control) deal with this all the time. The common solutions are:
1 – Swinger bypass
If a sensor crosses the threshold repeatedly, it’s put into bypass until reset. Settings are typically:
Only send X alarms in Y minutes, then stop until either a user reset or a set cooldown time. (Hydros uses the cooldown option, but somewhat crudely).
3 – Escalation rules
Send the first alarm, then suppress repeats unless the condition clearly worsens (crosses a higher threshold or persists beyond a time limit).
These (3) concepts are often used together or in hybrid combinations, depending on the system.
I agree with the above regarding the GHL alarms. There needs to be some added logic and information. While these can be somewhat scripted with email macros, the device notifications are lacking. Honestly, I can't even really get them to work correctly because the app is not always live in the background, SMS is long deprecated, etc. I would love to see this fully reworked.
The problem is well known in automation and monitorin: when a sensor hovers around a setpoint, you get endless nuisance alarms. Hysteresis doesn’t solve the problem. Other industries (fire, security, process control) deal with this all the time. The common solutions are:
1 – Swinger bypass
If a sensor crosses the threshold repeatedly, it’s put into bypass until reset. Settings are typically:
- Trigger count + trigger window: “If this alarm trips X times in Y minutes, bypass it.”
- Reset options: reset after a fixed period if swings stop, reset after a fixed period regardless, or require manual reset.
Only send X alarms in Y minutes, then stop until either a user reset or a set cooldown time. (Hydros uses the cooldown option, but somewhat crudely).
3 – Escalation rules
Send the first alarm, then suppress repeats unless the condition clearly worsens (crosses a higher threshold or persists beyond a time limit).
These (3) concepts are often used together or in hybrid combinations, depending on the system.
I agree with the above regarding the GHL alarms. There needs to be some added logic and information. While these can be somewhat scripted with email macros, the device notifications are lacking. Honestly, I can't even really get them to work correctly because the app is not always live in the background, SMS is long deprecated, etc. I would love to see this fully reworked.
