scrutiny

Scrutiny

Hard drive health dashboard powered by SMART data. Scrutiny runs smartctl against your drives on a schedule, stores historical metrics in a bundled InfluxDB instance, and surfaces failures — both raw SMART attribute failures and Backblaze-derived failure statistics — in a clean web UI.

This deployment uses the omnibus image, which bundles the web/API server, the drive collector (cron-scheduled), and InfluxDB into a single container. No separate database container is needed.

Architecture

┌──────────────────────────────────────────┐
│  scrutiny (omnibus container)            │
│                                          │
│  ┌────────────┐  ┌──────────────────┐   │
│  │ collector  │  │ web / API :8080  │   │
│  │ (cron)     │  │                  │◄──┼── Traefik → browser
│  └─────┬──────┘  └────────┬─────────┘   │
│        │                  │             │
│        └──────┬───────────┘             │
│               ▼                         │
│        ┌────────────┐                   │
│        │  InfluxDB  │                   │
│        └────────────┘                   │
└──────────────────────────────────────────┘
         │
  /dev/sda, /dev/sdb, ...  (host drives passed in)

The collector runs on a schedule (default: 6 hours) and sends SMART data to the internal InfluxDB. The web UI reads from InfluxDB and serves the dashboard.

Prerequisites

1. Create storage directories on the host

mkdir -p /mnt/SSD/Containers/scrutiny/config
mkdir -p /mnt/SSD/Containers/scrutiny/influxdb

2. Identify your drives

On the TrueNAS host, list all physical disks:

lsblk -d -o NAME,TYPE,SIZE,MODEL | grep disk

Or with smartctl:

smartctl --scan

Note the device paths (e.g. /dev/sda, /dev/nvme0n1) — you'll need to list them in compose.yaml.

3. Update the devices list in compose.yaml

Edit the devices: block to match your actual drives:

devices:
  - /dev/sda
  - /dev/sdb
  - /dev/sdc
  - /dev/nvme0n1
  # add every drive you want monitored

Scrutiny will only see drives explicitly listed here. Drives not listed are silently ignored.

Quick Start

cp example.env .env
# Edit .env — set SCRUTINY_DOMAIN at minimum
docker compose up -d

Verify

docker compose ps
docker compose logs -f

Wait 30–60 seconds for InfluxDB to initialize, then navigate to https://<SCRUTINY_DOMAIN>. The first SMART scan runs immediately on container start; subsequent scans follow the 6-hour schedule.

Configuration

.env variables

Variable	Default	Description
TZ	America/New_York	Timezone (controls cron schedule display)
SCRUTINY_DOMAIN	—	Hostname Traefik routes to Scrutiny
SCRUTINY_PORT	8096	Host port (direct LAN access only — Traefik ignores this)

scrutiny.yaml (optional advanced config)

For advanced configuration (collector schedule, notification webhooks, alert thresholds), drop a scrutiny.yaml in /mnt/SSD/Containers/scrutiny/config/. Scrutiny reads it on startup.

Example — change the collector schedule to every 12 hours:

# /mnt/SSD/Containers/scrutiny/config/scrutiny.yaml
collector:
  cron: "0 */12 * * *"

Full config reference: https://github.com/AnalogJ/scrutiny/blob/master/example.scrutiny.yaml

Capabilities

Scrutiny requires two Linux capabilities to read SMART data:

Capability	Required for
SYS_RAWIO	SATA/SAS ATA passthrough (most spinning and SATA SSD drives)
SYS_ADMIN	NVMe IOCTL passthrough (NVMe SSDs)

These are granted via cap_add in compose.yaml — not full privileged: true. If you have no NVMe drives, SYS_ADMIN can be removed from the compose file.

Storage

Data	Host path
Scrutiny config	/mnt/SSD/Containers/scrutiny/config
InfluxDB time-series data	/mnt/SSD/Containers/scrutiny/influxdb

Notifications

Scrutiny can push notifications to Slack, Discord, PagerDuty, email, and others via shoutrrr. Configure via scrutiny.yaml:

notify:
  urls:
    - "discord://token@webhookid"
    - "slack://tokenA/tokenB/tokenC"

Notifications fire when a drive's SMART status changes (passes → failed or failed → passes).

Maintenance

Force an immediate SMART scan

docker exec scrutiny scrutiny-collector-metrics run

Update image

docker compose pull
docker compose up -d

Check InfluxDB health

docker exec scrutiny influx ping