Know when a sensor goes quiet

HydroNode learns how often each sensor reports and tells you when it is late or gone. Set up delayed and offline alerts, push and email, and automations that react.

HydroNode 6 min read
The monitoring settings of a HydroNode sensor with interval, thresholds and alert channelsThe monitoring settings of a HydroNode sensor with interval, thresholds and alert channels

The nasty thing about a dead sensor: its chart looks calm. No spikes, no warnings, just nothing. Until you notice that the last value is three days old. HydroNode makes sure that does not happen to you, and learns the rhythm of every single sensor to do it.

Four states, one rule

Every sensor is Healthy, Delayed, Offline or has no data yet. Fleet, the sensor list, the widgets, automations and the alert email all read this one rule, so they always agree.

  1. HealthyReadings arrive on time
  2. DelayedAfter 3 missed intervals
  3. OfflineAfter 12 missed intervals
  4. Back onlineFirst reading after the outage
The thresholds are multiples of the interval and move with it when the rhythm changes.
Learned intervalFrom the gaps between recent readings, as a median. Outliers do not throw it off.
Delayed after 3×Three intervals without a value. Never sooner than 5 minutes.
Offline after 12×Twelve intervals without a value. Never sooner than 15 minutes.
Staying calmAt most one message every 30 minutes, unless it escalates from delayed to offline.

Set up monitoring

  1. Open the settingsSensor page, Settings, then Monitoring on the left. Right now at the top shows how the sensor is doing.
  2. Check the intervalUnder Reporting interval you see what HydroNode learned, for example Expected every 30 minutes. If your device sends irregularly, switch to Set by me and enter the real value.
  3. Adjust the thresholdsUnder Thresholds, change the multiple for delayed and offline with − and +. A device in deep sleep that sometimes misses a round is happier with a higher offline multiple.
  4. Pick the channelsUnder Tell me, switch push and email on or off per stage, plus Back online and Device faults.
Monitoring settings with the current state, learned interval and thresholdsMonitoring settings with the current state, learned interval and thresholds
The current state with a timeline at the top, the learned interval and both thresholds below.
Defaults Delayed stays silent, offline sends a push, and back online only arrives if an alert actually went out before. The first check after a sensor is created never reports anything, so a fresh sensor does not alarm right away.

Muted means muted

If a sensor is set to Off under SettingsNotifications, that covers monitoring pushes too. Handy for a test rig that gets unplugged a lot. The app still shows its state.

Notification settings of a sensor with Off, Important only and EverythingNotification settings of a sensor with Off, Important only and Everything
Three levels per sensor. The choice is yours alone and syncs to all your devices.

More than a message: automations

Sometimes a quiet sensor should make something happen, not just send a note. Create a rule for that under OperationsAutomations.

  1. New automationGive it a name and pick the sensor.
  2. TriggerUnder Trigger pick The sensor goes quiet, then Is delayed, Goes offline or Is back online.
  3. ActionSend a notification, Send a command to the device with a saved command template, or Call a webhook with a signed POST to your own HTTPS address.
  4. BrakeUnder Run at most set how often the rule may fire.
The form for a new automation with trigger and actionThe form for a new automation with trigger and action
A rule watches one sensor and does exactly one thing when its condition is met.

A real-world example: when the greenhouse sensor goes offline, a webhook tells your home automation to switch the irrigation to a fixed schedule until readings come back.

Read more