Automations

With the integration connected, Home Assistant can look after your containers by itself: tell you when a new version is out, update the containers you choose at night, warn you when one stops. The examples on this page are ready to paste.

What an automation can do

The integration has no actions of its own: its entities work with the actions Home Assistant already has.

To	Action	Entity
Update one container	`update.install`	The Update entity of the container, such as `update.jellyfin_update`
Update every container	`button.press`	Update all, `button.d2ha_update_all`
Look for new versions now	`button.press`	Check for updates, `button.d2ha_check_for_updates`
Delete the unused images	`button.press`	Clean images, `button.d2ha_clean_images`

An automation can start from the State sensor of a container, such as sensor.jellyfin_state, or from the Updates available counter, sensor.d2ha_updates_available. What you get lists every entity with its attributes.

Add an example

In Home Assistant, open Settings → Automations & scenes and press Create automation, then Create new automation.
Open the menu ⋮ and choose Edit in YAML.
Paste the example in place of what is there, put in your entity ids and Save.

Your entity ids

The examples use the ids of a Home Assistant in English and of two containers called jellyfin and vaultwarden. Yours depend on the language and on the names of your containers: find them in Settings → Devices & services → D2HA. Replace notify.notify too, with your own notification action: for the phone app it is notify.mobile_app_ followed by the name of the phone.

A message when new updates appear

alias: "D2HA: new container updates"
triggers:
  - trigger: state
    entity_id: sensor.d2ha_updates_available
    not_from:
      - unknown
      - unavailable
conditions:
  - condition: template
    value_template: >-
      {{ trigger.from_state is not none and
         trigger.to_state.state | int(0) > trigger.from_state.state | int(0) }}
actions:
  - action: notify.notify
    data:
      title: Container updates
      message: >-
        {{ state_attr('sensor.d2ha_updates_available', 'containers') | join(', ') }}

The message leaves every time the number of updates grows, and names the containers that are waiting. It does not leave when the number goes down after an update, or when Home Assistant or D2HA restarts.

The number counts every container of D2HA, while the names are only those of the containers you show to Home Assistant.

Update the containers you choose at night

alias: "D2HA: update chosen containers at night"
triggers:
  - trigger: time
    at: "04:00:00"
actions:
  - action: update.install
    target:
      entity_id: >-
        {{ ['update.jellyfin_update', 'update.vaultwarden_update']
           | select('is_state', 'on') | list }}

Every night at 4:00, the containers of the list that have an update are updated, one after the other. With nothing to update, nothing happens.

Keep the filter

select('is_state', 'on') keeps the containers that have an update. Without it, a container of the list that is already up to date ends the run with the error No update available, even though the other containers are updated.

Put in the list only containers that have Install in Home Assistant. D2HA cannot update a container whose image was built on your machine, and updates nothing in read-only mode.

Leave out the breaking changes

D2HA marks an update when its release notes have a Breaking changes section. To keep those for yourself, use this action in place of the one above:

actions:
  - action: update.install
    target:
      entity_id: >-
        {{ ['update.jellyfin_update', 'update.vaultwarden_update']
           | select('is_state', 'on')
           | reject('is_state_attr', 'breaking_changes', true) | list }}

Update everything at night

alias: "D2HA: update everything at night"
triggers:
  - trigger: time
    at: "04:00:00"
actions:
  - action: button.press
    target:
      entity_id: button.d2ha_update_all

Update all updates every container that has an update, one at a time: also the containers you hide from Home Assistant, and also the updates with breaking changes. It never updates D2HA itself or its socket proxy. With nothing to update, nothing happens.

A message when a container stops

alias: "D2HA: a container stopped"
triggers:
  - trigger: state
    entity_id:
      - sensor.jellyfin_state
      - sensor.vaultwarden_state
    to:
      - exited
      - dead
    for: "00:02:00"
actions:
  - action: notify.notify
    data:
      title: Container stopped
      message: >-
        {{ device_attr(trigger.entity_id, 'name') }}:
        {{ state_attr(trigger.entity_id, 'status') }}

The message leaves when one of the containers has been stopped for two minutes, and says what Docker reports about it, such as Exited (137) 2 minutes ago. The two minutes leave out a restart and an update, where the container is stopped only for a moment. When Home Assistant loses D2HA and finds it again, as after a restart of D2HA, a container that is still stopped is announced again.

An unhealthy container

A container with a health check can be unhealthy while it still runs: its State sensor stays on running, and the health is in the health attribute. This second automation tells you when a container has been unhealthy for two minutes:

alias: "D2HA: a container is unhealthy"
triggers:
  - trigger: state
    entity_id:
      - sensor.jellyfin_state
      - sensor.vaultwarden_state
    attribute: health
    to: unhealthy
    for: "00:02:00"
actions:
  - action: notify.notify
    data:
      title: Container unhealthy
      message: "{{ device_attr(trigger.entity_id, 'name') }}"

Without Home Assistant

D2HA can also send these messages by itself: new versions, containers that stop with an error and unhealthy ones. See Telegram.

Clean the images every week

alias: "D2HA: clean images every week"
triggers:
  - trigger: time
    at: "05:00:00"
conditions:
  - condition: time
    weekday:
      - sun
actions:
  - action: button.press
    target:
      entity_id: button.d2ha_clean_images

Every Sunday at 5:00, D2HA deletes the images that no container uses, running or stopped, as Clean images does in the web interface.

When an automation does not run

Open the automation and press Traces: every run is there step by step, with the error when one stopped it. update.install waits for the update to end: if it fails, D2HA puts the old container back and the run ends with the error.

Something wrong or missing on this page? Open an issue.