GitHub → Install
v2.4.2 · Marketplace & HACS Default

A health score for
your Home Assistant.

HAGHS checks system load, your database, unavailable entities, updates, repairs and backups, then gives you one honest 0–100 score and a short list of exactly what to fix. All local. All open.

Install HAGHS View on GitHub
  • 100% local
  • Apache-2.0
  • No outbound traffic
HAGHS example
97/100
Healthy
Hardware
100 /100
Application
95 /100
  • 1 update pendingESPHome · counted after 7 days pending −5 pts
  • 1 open repairSettings · System · Repairs −5 pts
  • Bonus · set a recorder entity filterConfig audit: retention bonus already earned +5 pts
sensor.system_ha_global_health_score local
Featured in

Why HAGHS

Stays on your hardware

No cloud, no telemetry, no outbound traffic. HAGHS reads kernel pressure, your database and Home Assistant's own registries directly. Nothing about your home leaves your network.

Nothing hidden

Every threshold and penalty is in the source and in the docs. No surprise weights, no black box. Read the math, argue with it on GitHub, or send a pull request.

Tells you what to fix

Every missing point comes with a plain-language recommendation, in your Home Assistant language. The dashboard cards link straight to the right screen.

// the score

Two pillars.
One number.

The index weights long-term hygiene over short load spikes. Floor rounding means a perfect 100 is only reached by systems with nothing left to fix.

Score = ⌊ Hardware × 0.4 + Application × 0.6 ⌋

Zombies, open repairs and overdue watched entities cap the application score at 99, so a bonus can never hide a real problem.

Hardware · 40%

  • CPU, RAM and I/O via kernel PSI
  • Sensor fallback without PSI
  • Disk space per storage type
  • Raspberry Pi under-voltage

Application · 60%

  • Zombie entities
  • Updates and Core age
  • Integration health, repairs
  • Database, backups, config audit
// advisor

The kind of things
you'd never spot.

Examples of what the advisor reports. Each finding names the metric, the points it costs and where to look in Home Assistant.

// zombiesup to −20

7 entities unavailable past the grace period, across sensor, switch and light. Each one says why, and only config_entry_missing hints at a real orphan.

→ Settings · Entities
// repairs−5 each

2 open repairs from other integrations. Home Assistant's own Ignore takes one out of the score.

→ Settings · System · Repairs
// hardware · psiCPU −50

CPU stall time at 34%: tasks are waiting, not just busy. Classic utilisation sensors would miss this.

→ Check add-ons and integrations
// database−10

Database at 1.9 GB, over the 1.5 GB limit for 200 entities. Above 2.5 times the limit it costs 30.

→ Recorder retention and filters
// backups−30

Home Assistant reports your backups as stale. One of the largest single deductions in the application pillar.

→ Settings · System · Backups
// config auditup to +10

Bonus for recorder hygiene: history kept at most 10 days (+5) and a user-set entity filter (+5).

→ Recorder configuration

Every rule, threshold and cap: docs/scoring.md

  • Marketplace · HACS
    Default repository, one-click install
  • v2.4.2 · Apache-2.0
    latest release, open licence
  • 5,000+ installs
    counted by HA's opt-in analytics
  • PSI-aware
    CPU · RAM · I/O stall time on Linux
// install

Up and scoring
in five minutes.

  1. InstallHome Assistant 2026.11 and later: Settings › Marketplace, search “HAGHS”, Install. Up to 2026.10: download HAGHS in HACS and restart Home Assistant.
  2. Add the integrationSettings › Devices & Services › Add Integration › “HAGHS”.
  3. Follow the setupStorage type, ignore labels and patterns. CPU and RAM sensors are only needed when your host has no PSI. Database size and disk usage are detected automatically.
  4. Add a dashboard cardReady-made Lite and Pro cards show the score, the breakdown and links to every fix.

Getting started guide →

automations.yaml · example
# Get a notification when the score drops below 80
- alias: "HAGHS: health score dropped"
  triggers:
    - trigger: numeric_state
      entity_id: sensor.system_ha_global_health_score
      below: 80
  actions:
    - action: notify.notify
      data:
        title: "Home Assistant health: {{ states('sensor.system_ha_global_health_score') }}"
        message: "{{ state_attr('sensor.system_ha_global_health_score', 'recommendations') }}"

See your score in five minutes.

Free, local and open source. Questions and ideas are welcome in the GitHub discussions.

Install HAGHS Join the discussion