112 lines
3.5 KiB
Markdown
112 lines
3.5 KiB
Markdown
# DB Monitor
|
|
|
|
> Monitor your PostgreSQL databases without the Prometheus + Grafana overhead.
|
|
|
|
A lightweight, self-hosted monitoring tool built with Go. Point it at your PostgreSQL instances, and it gives you a real-time dashboard with metrics, charts, and Telegram alerts — out of the box.
|
|
|
|
No config files to wrestle with. No YAML. Just set your connection string and go.
|
|
|
|
## Getting Started
|
|
|
|
### Option 1 — Docker Hub (recommended)
|
|
|
|
```bash
|
|
docker run -d \
|
|
--name db-monitor \
|
|
-p 9432:9432 \
|
|
-v db-monitor-data:/data \
|
|
-e DB_MONITOR_DSN="postgresql://user:pass@host:5432/mydb" \
|
|
mahmoudmo123/db-monitor:latest
|
|
```
|
|
|
|
Dashboard will be at `http://localhost:9432`.
|
|
|
|
### Option 2 — docker compose
|
|
|
|
```bash
|
|
git clone https://github.com/mahmoud/db-monitor.git
|
|
cd db-monitor
|
|
cp .env.example .env # add your PostgreSQL DSN here
|
|
docker compose up -d
|
|
```
|
|
|
|
Dashboard will be at `http://localhost:9432`.
|
|
|
|
### Option 3 — Ghaymah Cloud
|
|
|
|
1. Go to **deploy.ghaymah.systems** → **Deploy App**
|
|
2. **Container image:** `mahmoudmo123/db-monitor:latest`
|
|
3. **Port:** `9432`
|
|
4. **Environment variables** (optional):
|
|
|
|
| Key | Value |
|
|
|---|---|
|
|
| `DB_MONITOR_PORT` | `9432` |
|
|
| `DB_MONITOR_INTERVAL` | `15` |
|
|
| `DB_MONITOR_SQLITE_PATH` | `/data/db-monitor.db` |
|
|
|
|
5. **Volume:** mount `/data` for persistent storage
|
|
6. Click **Deploy App** ✅
|
|
|
|
> **Adding databases:** No need to set `DB_MONITOR_DSN` upfront. Once deployed, open the dashboard → **Settings** and add your PostgreSQL connections from the UI.
|
|
|
|
## What It Tracks
|
|
|
|
- **Connections** — active, idle, total, max, usage percentage
|
|
- **Queries** — currently running, long-running (>5min), waiting
|
|
- **Locks** — active, waiting, deadlocks
|
|
- **Performance** — cache hit ratio, commits/rollbacks, tuple operations
|
|
- **Replication** — lag, primary/replica status
|
|
- **Size & uptime** — database size, PostgreSQL version
|
|
|
|
## Alerting
|
|
|
|
If you configure Telegram, you get notified when things go wrong:
|
|
|
|
- Database goes **down**
|
|
- Connection usage exceeds **80%**
|
|
- A query runs longer than **5 minutes**
|
|
- **Deadlocks** detected
|
|
- More than **5 locks** waiting
|
|
- Cache hit ratio drops below **90%**
|
|
|
|
## Permissions
|
|
|
|
DB Monitor only needs read access. Create a dedicated user:
|
|
|
|
```sql
|
|
CREATE USER db_monitor WITH PASSWORD 'your_password';
|
|
GRANT CONNECT ON DATABASE your_db TO db_monitor;
|
|
GRANT pg_monitor TO db_monitor;
|
|
```
|
|
|
|
## How It Works
|
|
|
|
```
|
|
┌─────────────────────────────────┐
|
|
│ DB Monitor │
|
|
│ │
|
|
│ Collector ──→ SQLite (metrics) │
|
|
│ │ │ │
|
|
│ │ Alert Engine │
|
|
│ │ (Telegram) │
|
|
│ ↓ ↓ │
|
|
│ REST API + Web Dashboard │
|
|
│ :9432 │
|
|
└────────────┬────────────────────┘
|
|
│
|
|
┌──────────┼──────────┐
|
|
↓ ↓ ↓
|
|
PostgreSQL PostgreSQL PostgreSQL
|
|
```
|
|
|
|
Every N seconds (default: 15), the collector queries `pg_stat_activity`, `pg_stat_database`, `pg_locks`, and friends. Metrics get stored in an embedded SQLite database. The dashboard reads from SQLite and shows live charts via Chart.js.
|
|
|
|
## Built With
|
|
|
|
Go · SQLite · Chart.js · Docker
|
|
|
|
## License
|
|
|
|
MIT
|