Monitoring partial backup failures : no event-driven signal that I can find?

Configuration
, , ,
1

Title: Monitoring partial backup failures - no event-driven signal that I can find?

Context

I want an automation that alerts me when a backup goes partial
where the backup file gets written, but the supervisor failed to include
one or more add-ons (e.g. an add-on's Docker image fails checksum during
docker save, so it gets skipped). The UI clearly flags this:

"This backup was not created successfully. Some data is missing.
Failed apps: <addon name>"

The backup is technically usable, but it's not what was requested, and
silently degraded backups are the worst kind. I'd like to know in HA,
not by clicking into the backup detail in the UI.

By all rights, a Repair action is what I would expect on any failure.

What I observed

The event.backup_automatic_backup entity dumps as:

state: 2026-05-18T18:36:01.286+00:00
attributes:
backup_stage: null
event_type: completed
event_types: ["completed", "failed", "in_progress"]
failed_reason: null
friendly_name: Backup Automatic backup

So the entire vocabulary is completed / failed / in_progress. On
a partial backup, event_type is completed. No failed_reason,
no attribute listing skipped add-ons, no failed_addons / failed_agents /
failed_folders despite those keys appearing in the integration's
Repairs translation strings (automatic_backup_failed_addons etc).

What I tried

  1. Trigger on event.backup_automatic_backup with attribute: event_type, to: failed —
    correct for hard failures but blind to partials.

  2. Read failed_addons / failed_agents / failed_folders attributes
    on the entity — those attributes don't exist on my instance.

  3. Trigger on repairs_issue_registry_updated event in case the
    integration files a Repairs issue for partials. Tested both:

    • Manual partial backup: no Repairs issue, no event.
    • Automatic partial backup (triggered via backup.create_automatic
      to mimic the scheduler): same — no Repairs issue, no event.

    Listening on * in Developer Tools → Events during a partial backup,
    nothing fires that I can hook to.

Current workaround

Daily shell_command automation that polls the supervisor API
(/backups, /backups/<slug>/info, /addons), diffs installed add-ons
against what's in the most recent backup, and alerts if anything installed
is missing. Catches it reliably, but is independent of any HA event surface
— purely polling. Happy to share the YAML if it's useful to others.

Versions

  • Installation methodHome Assistant OS
  • Core2026.4.2
  • Supervisor2026.05.0
  • Operating System17.2
  • Frontend20260325.7

Questions

  1. Did I miss something? Is there an attribute, event, service
    response, or sensor anywhere that exposes partial-failure detail without
    polling the supervisor API directly?

  2. Are the automatic_backup_failed_addons Repairs strings reachable?
    They exist in strings.json but I can't find a path that fires them.
    Wrong test, gated behind a config option, or dead code?

  3. Has anyone built clean partial-failure monitoring on a recent HA
    version? Curious if there's a pattern I'm not aware of.

  4. Any open feature requests for partial-failure surfacing that I
    should "like" rather than file a new one? Closest I found:

    Neither specifically asks for a partial event_type or
    failed_addons attribute on the event entity. If nobody's filed that
    specifically, I'll open one — wanted to check here first.

Thanks!

1 Like