How Alarms Work
LoopString uses a two-tier alarm system with hysteresis to prevent alert flapping:
- Warning — Yellow indicator, informational notification
- Critical — Red indicator, urgent notification (email and/or SMS)
Each sensor can have independent high and low thresholds for both tiers.
Configuring Thresholds
Click any SensorStat card on the dashboard to open the inline threshold editor:
- Enable alarms for the sensor
- Set warning high and warning low values
- Set critical high and critical low values
- Configure hysteresis (dead-band) to prevent rapid on/off toggling
Hysteresis
Hysteresis adds a buffer zone around each threshold:
- A threshold is breached when the reading crosses it
- The alarm clears only when the reading returns past the threshold by the hysteresis amount
- Example: Critical high at 30 degrees with 2 degree hysteresis clears when reading drops below 28 degrees
This prevents alarms from rapidly toggling when a reading hovers near a threshold.
Alarm States
The alarm system uses a finite state machine:
- Normal — Reading within all thresholds
- Warning — Reading crossed warning threshold
- Critical — Reading crossed critical threshold
- Acknowledged — User acknowledged the alert (stops repeated notifications)
Notifications
When an alarm triggers:
- In-app notification — Appears in the notification center (bell icon) with unread badge
- Email — Sent if enabled in notification preferences
- SMS — Available on paid tiers for critical alarms
Configure notification preferences in Account Settings > Notifications.
Alert Events
Every alarm trigger creates an alert event stored in Firestore:
- Timestamp of breach
- Sensor and device identification
- Threshold that was crossed
- Reading value at time of breach
- Acknowledgment status and timestamp
View alert history in the Activity Timeline or the notification center.
Edge Alerting
Alarm evaluation runs in two places:
- Cloud Function — Monitors RTDB sensor values and triggers notifications
- Pi (optional) — Alert configs are mirrored to RTDB for local edge alerting, enabling responses even without cloud connectivity
Alarm Colors on Cards
Alarm status is visually tied to the SensorStat card:
- Card border color changes to match the alarm severity
- Colors respect unit conversion (thresholds display in your preferred units)
- Dashboard dark mode uses high-contrast alarm colors for visibility
Next Steps
Escalation Chains
For alerts that need multi-tier routing — escalating to different contacts if the first responder does not acknowledge — see Alert Escalation Chains. Escalation chains are assigned per alert and require a Pro tier subscription.