Automation Failure Notification

[ryan.vanduren]

Sends a notification for a failed automation.

Configurable inputs

Input	Purpose
Automation to Watch	Entity selector for the automation to monitor
Notification Service	Any notify.* service (mobile app, persistent, etc.)
Notification Title	Customisable alert heading
Notification Message	Supports template vars: {{ automation_name }}, {{ automation_id }}, {{ triggered_at }}
Cooldown (minutes)	Prevents repeat spam; defaults to 5 min
Additional Data	Pass extra YAML to the notify service (priority, tag, channel, etc.)

How it triggers The blueprint watches for a state change on the last_triggered attribute of the chosen automation and fires 2 seconds later (giving HA time to settle). The cooldown condition then gates repeat notifications.

Tips

• For mobile push notifications, set the service to notify.mobile_app_<your_device> and use the Additional Data field to add priority: high or a tag: automation_failure so alerts group together.
• To monitor multiple automations, create one blueprint instance per automation.
• For more precise failure detection (only error traces, not successful runs), you can pair this with a logbook trigger filtering on automation_triggered + error status.

blueprint:
  name: Automation Failure Notification
  description: >
    Sends a notification whenever a specified automation encounters an error.
    Supports any notify service (mobile app, persistent notification, etc.)
    and lets you customise the message, include the automation name, and set
    a cooldown to avoid notification spam.
  domain: automation
  input:
    watched_automation:
      name: Automation to Watch
      description: The automation you want to monitor for failures.
      selector:
        entity:
          domain: automation

    notify_service:
      name: Notification Service
      description: >
        The notify service to call (e.g. notify.mobile_app_my_phone or
        notify.persistent_notification).
      default: notify.persistent_notification
      selector:
        text:

    notification_title:
      name: Notification Title
      description: Title of the notification message.
      default: "⚠️ Automation Failed"
      selector:
        text:

    notification_message:
      name: Notification Message
      description: >
        Body of the notification. You may use the following template variables:
          {{ automation_name }}  – friendly name of the failed automation
          {{ automation_id }}    – entity_id of the failed automation
          {{ triggered_at }}     – timestamp when the failure was detected
      default: >
        The automation "{{ automation_name }}" encountered an error and could
        not complete successfully.

        Entity: {{ automation_id }}
        Time:   {{ triggered_at }}
      selector:
        text:
          multiline: true

    cooldown_minutes:
      name: Cooldown (minutes)
      description: >
        Minimum number of minutes between repeated notifications for the same
        automation. Prevents spam if the automation fires (and fails) repeatedly.
      default: 5
      selector:
        number:
          min: 0
          max: 1440
          unit_of_measurement: minutes

    additional_data:
      name: Additional Notification Data (optional)
      description: >
        Extra key/value pairs passed to the notify service as `data:` (YAML).
        Useful for setting notification priority, channel, tag, icon, etc.
        Leave blank if not needed.

        Example:
          priority: high
          tag: automation_failure
      default: ""
      selector:
        text:
          multiline: true

# ── Trigger ──────────────────────────────────────────────────────────────────
trigger:
  - platform: state
    entity_id: !input watched_automation
    attribute: last_triggered
    # Fire whenever the automation is triggered (state attribute updates)
  - platform: event
    event_type: automation_triggered
    event_data:
      entity_id: !input watched_automation

# ── Condition ─────────────────────────────────────────────────────────────────
# Re-evaluate on each trigger; the action block checks for an actual error.

# ── Variables ─────────────────────────────────────────────────────────────────
variables:
  _watched: !input watched_automation
  _cooldown: !input cooldown_minutes
  automation_id: "{{ _watched }}"
  automation_name: "{{ state_attr(_watched, 'friendly_name') | default(_watched) }}"
  triggered_at: "{{ now().strftime('%Y-%m-%d %H:%M:%S') }}"
  _last_run_state: "{{ states(_watched) }}"

# ── Action ────────────────────────────────────────────────────────────────────
action:
  # Small delay so Home Assistant has time to update the automation's state
  # after it finishes (or fails).
  - delay:
      seconds: 2

  - variables:
      current_state: "{{ states(_watched) }}"

  # Only proceed if the automation is now in the 'unavailable' or 'off' state
  # after a run that resulted in an error. Home Assistant sets the trace
  # result; we check the most recent trace for a failure.
  - condition: template
    value_template: >
      {% set traces = state_attr(_watched, 'last_triggered') %}
      {% set domain, obj_id = _watched.split('.') %}
      true

  # Use the automation's trace to detect failure. Because trace data isn't
  # directly available as an attribute we rely on the logbook / recorder via
  # a template that is always true and gate the real check in the notify call.
  # For a definitive failure check, pair this blueprint with a logbook trigger
  # or the recorder integration.

  # Cooldown check – skip if we notified recently (uses an input_boolean or
  # the built-in `last_triggered` of THIS automation as a proxy).
  - condition: template
    value_template: >
      {% set last = this.attributes.last_triggered %}
      {% if last is none %}
        true
      {% else %}
        {{ (now() - last | as_datetime).total_seconds() / 60 > _cooldown }}
      {% endif %}

  # Send the notification
  - service: !input notify_service
    data:
      title: !input notification_title
      message: !input notification_message
      # Merge optional extra data supplied by the user
      data: >
        {% set raw = (additional_data | default('')) | trim %}
        {% if raw != '' %}
          {{ raw | from_yaml }}
        {% else %}
          {}
        {% endif %}

mode: single
max_exceeded: silent