clawdia/Home-Assistant

Fork 0

Files

clawdia bdd5b1bfda Add full infrastructure documentation: cleanup, InfluxDB migration, Grafana dashboards

2026-07-01 23:42:24 +00:00

6.5 KiB

Raw Blame History

Home Assistant Infrastructure Documentation

Last updated: 2026-July
HA Version: 2026.3.1 (update available: 2026.6.4)
HA Host: 10.10.0.123

Infrastructure Overview

┌─────────────────────┐       ┌──────────────────────────┐       ┌──────────────────────┐
│   Home Assistant    │──────▶│       InfluxDB 2.7       │──────▶│       Grafana        │
│   10.10.0.123       │       │    SRvDocker1 (10.10.0.41)│       │   10.10.0.41:3001     │
│   HA 2026.3.1       │       │    port 8086              │       │   v13.1.0             │
└─────────────────────┘       └──────────────────────────┘       └──────────────────────┘

Home Assistant pushes time-series metrics to InfluxDB v2.7
InfluxDB stores metrics in the home_assistant bucket (13-month retention)
Grafana reads from InfluxDB and renders dashboards
Both InfluxDB and Grafana run as Docker containers on SRvDocker1 (10.10.0.41)

1. Cleanup Summary

The Home Assistant host disk was at 83% usage and is now at 46% after cleanup.

Action	Space Freed
Old HA backups (versions pre-2026)	~28 GB
Orphaned temporary directories	~7.2 GB
Total	~35.2 GB

Backups older than 3 months were identified and removed.
Orphaned temp directories from previous addons/upgrades were cleaned.
HA recorder database was compacted via auto-repair.

2. InfluxDB 2.7 Migration

Background

The old InfluxDB addon (installed inside HA) had been idle since December 2025.
It was removed entirely to free resources on the HA host.
A standalone InfluxDB 2.7.12 instance was deployed on SRvDocker1.

Deployment Details

Field	Value
Version	InfluxDB 2.7.12
Host	SRvDocker1 (10.10.0.41)
Port	8086
Docker Compose	`/home/adminx/influxdb/docker-compose.yml`
Bucket	`home_assistant`
Organization	`swordrock`
Retention	13 months (9456 hours)
API Version	v2
SSL	Off (internal network)
Admin Token	Stored in locky Vaultwarden

Docker Compose

# /home/adminx/influxdb/docker-compose.yml
services:
  influxdb:
    image: influxdb:2.7.12
    container_name: influxdb
    restart: unless-stopped
    ports:
      - "8086:8086"
    volumes:
      - /data/influxdb:/var/lib/influxdb2
      - /etc/influxdb2:/etc/influxdb2
    environment:
      - INFLUXDB_REPORTING_DISABLED=true

Home Assistant Integration

Method: UI-based setup (Settings → Devices & Services → InfluxDB)
Why UI and not YAML? The influxdb: YAML configuration did not work on HA 2026.3.1. The UI-based integration was the only method that successfully connected and started pushing data.
YAML cleanup: The influxdb: block was removed from configuration.yaml to avoid conflicts.

3. Recorder Tuning

To reduce SQLite DB writes and I/O pressure on the HA host, recorder settings were added to configuration.yaml.

Configuration

recorder:
  purge_keep_days: 7
  commit_interval: 30

Setting	Value	Purpose
`purge_keep_days`	7	Keep only 7 days of history in the SQLite DB
`commit_interval`	30	Batch commits every 30 seconds (reduces write frequency)

The full configuration snippet is available in configuration-snippets.yaml.

Rationale

Historical data is now stored in InfluxDB (13-month retention), so the HA SQLite DB only needs a short window.
Lower commit frequency reduces disk I/O spikes.
This keeps the HA SQLite DB small and fast while InfluxDB handles long-term storage.

4. Grafana Dashboards

Deployment

Field	Value
Version	Grafana 13.1.0
Host	SRvDocker1 (10.10.0.41)
Port	3001
Login	`admin`
Password	Stored in locky Vaultwarden ("Grafana Admin" item)
Data Source	InfluxDB (Flux queries against `home_assistant` bucket)

Dashboards

🏠 Home Energy

Solar production (Watts, daily/monthly totals)
Grid consumption (import/export, net metering)
Voltage (L1, L2, L3 phases)
Frequency (grid frequency monitoring)

🌡️ Climate

Indoor/outdoor temperature (all sensors)
Humidity (indoor rooms, outdoor)
Weather (forecast vs actual, pressure, wind)

📷 Security Cameras

Camera status (online/offline)
Motion event counts
Stream health and uptime

5. Known Issues

Issue	Status	Notes
HA 2026.3.1 InfluxDB YAML config	⚠️ Workaround applied	YAML `influxdb:` block didn't work. UI-based integration is the only working method.
HA update pending	⚠️ Available	HA 2026.6.4 is available. Upgrade may change InfluxDB integration behavior — test after upgrade.
No network bandwidth data in HA	ℹ️ By design	UniFi traffic/bandwidth data flows to Zabbix, not HA. No plan to change.
No Zabbix → InfluxDB streaming	❌ Not configured	Zabbix metrics are not yet streamed to InfluxDB. Future enhancement: could pipe Zabbix alerts/metrics into InfluxDB for unified Grafana dashboards.

6. Maintenance Notes

Checking HA Disk

ssh ha-host df -h /

InfluxDB Container Health

ssh srvoc0 "ssh 10.10.0.41 docker ps | grep influxdb"

Grafana Container Health

ssh srvoc0 "ssh 10.10.0.41 docker ps | grep grafana"

InfluxDB CLI

docker exec -it influxdb influx query 'from(bucket: "home_assistant") |> range(start: -1h) |> limit(n: 5)'

Backup Reminders

HA backups auto-prune at 7 days — verify retention is working.
InfluxDB data lives on SRvDocker1 at /data/influxdb — ensure this is included in backup policy.
Grafana dashboard JSONs should be exported periodically.

ZWave-JS-UI Repository — Z-Wave network management
Zabbix Monitoring — Network monitoring, UniFi data
locky Vaultwarden — All credentials (InfluxDB admin token, Grafana admin password)

6.5 KiB Raw Blame History Unescape Escape