86 lines
3.0 KiB
Markdown
86 lines
3.0 KiB
Markdown
|
# Prometheus Metrics
|
||
|
|
|
||
|
|
Hemmelig provides a Prometheus-compatible metrics endpoint for monitoring your instance.
|
||
|
|
|
||
|
|
## Enabling Metrics
|
||
|
|
|
||
|
|
1. Go to **Dashboard > Instance > Metrics** tab
|
||
|
|
2. Enable the **Enable Prometheus Metrics** toggle
|
||
|
|
3. Optionally, set a **Metrics Secret** for authentication
|
||
|
|
4. Save the settings
|
||
|
|
|
||
|
|
## Endpoint
|
||
|
|
|
||
|
|
```
|
||
|
|
GET /api/metrics
|
||
|
|
```
|
||
|
|
|
||
|
|
## Authentication
|
||
|
|
|
||
|
|
If a metrics secret is configured, you must include it as a Bearer token in the `Authorization` header:
|
||
|
|
|
||
|
|
```bash
|
||
|
|
curl -H "Authorization: Bearer YOUR_METRICS_SECRET" https://your-instance.com/api/metrics
|
||
|
|
```
|
||
|
|
|
||
|
|
If no secret is configured, the endpoint is accessible without authentication (not recommended for production).
|
||
|
|
|
||
|
|
## Available Metrics
|
||
|
|
|
||
|
|
### Application Metrics
|
||
|
|
|
||
|
|
| Metric | Type | Description |
|
||
|
|
| ---------------------------------------- | --------- | -------------------------------------------- |
|
||
|
|
| `hemmelig_secrets_active_count` | Gauge | Current number of active (unexpired) secrets |
|
||
|
|
| `hemmelig_users_total` | Gauge | Total number of registered users |
|
||
|
|
| `hemmelig_visitors_unique_30d` | Gauge | Unique visitors in the last 30 days |
|
||
|
|
| `hemmelig_visitors_views_30d` | Gauge | Total page views in the last 30 days |
|
||
|
|
| `hemmelig_http_request_duration_seconds` | Histogram | Duration of HTTP requests in seconds |
|
||
|
|
|
||
|
|
### Default Node.js Metrics
|
||
|
|
|
||
|
|
The endpoint also exposes default Node.js runtime metrics including:
|
||
|
|
|
||
|
|
- `nodejs_heap_size_total_bytes` - Process heap size
|
||
|
|
- `nodejs_heap_size_used_bytes` - Process heap size used
|
||
|
|
- `nodejs_external_memory_bytes` - Node.js external memory
|
||
|
|
- `nodejs_eventloop_lag_seconds` - Event loop lag
|
||
|
|
- `nodejs_active_handles_total` - Number of active handles
|
||
|
|
- `nodejs_active_requests_total` - Number of active requests
|
||
|
|
- `process_cpu_user_seconds_total` - User CPU time spent
|
||
|
|
- `process_cpu_system_seconds_total` - System CPU time spent
|
||
|
|
- `process_start_time_seconds` - Process start time
|
||
|
|
- `process_resident_memory_bytes` - Resident memory size
|
||
|
|
|
||
|
|
## Prometheus Configuration
|
||
|
|
|
||
|
|
Add the following job to your `prometheus.yml`:
|
||
|
|
|
||
|
|
```yaml
|
||
|
|
scrape_configs:
|
||
|
|
- job_name: 'hemmelig'
|
||
|
|
scrape_interval: 30s
|
||
|
|
static_configs:
|
||
|
|
- targets: ['your-instance.com']
|
||
|
|
metrics_path: '/api/metrics'
|
||
|
|
scheme: https
|
||
|
|
# If using authentication:
|
||
|
|
authorization:
|
||
|
|
type: Bearer
|
||
|
|
credentials: 'YOUR_METRICS_SECRET'
|
||
|
|
```
|
||
|
|
|
||
|
|
## Grafana Dashboard
|
||
|
|
|
||
|
|
You can create a Grafana dashboard to visualize these metrics. Here's an example panel query for active secrets:
|
||
|
|
|
||
|
|
```promql
|
||
|
|
hemmelig_secrets_active_count
|
||
|
|
```
|
||
|
|
|
||
|
|
## Security Considerations
|
||
|
|
|
||
|
|
- Always use a strong, randomly generated secret for the metrics endpoint in production
|
||
|
|
- Consider using network-level restrictions (firewall, VPN) to limit access to the metrics endpoint
|
||
|
|
- The metrics endpoint does not expose any sensitive data (secret contents, user data, etc.)
|