nthlayer generate-dashboard¶

Generate Grafana dashboards from a service specification.

Synopsis¶

nthlayer generate-dashboard <service-file> [options]

Description¶

The generate-dashboard command creates Grafana dashboards tailored to your service's technology stack. Dashboards include SLO panels, health metrics, and technology-specific visualizations.

Options¶

Option	Description
`--output, -o PATH`	Output file path (default: `generated/dashboards/{service}.json`)
`--env, --environment ENV`	Environment name (dev, staging, prod)
`--auto-env`	Auto-detect environment from CI/CD context
`--dry-run`	Print dashboard JSON without writing file
`--full`	Include all template panels (default: overview only)
`--prometheus-url, -p URL`	Prometheus URL for metric discovery

Examples¶

Basic Generation¶

nthlayer generate-dashboard services/payment-api.yaml

Generates a Grafana dashboard JSON:

{
  "title": "payment-api - Service Dashboard",
  "uid": "payment-api-dashboard",
  "tags": ["nthlayer", "payment-api", "tier-1"],
  "panels": [
    {
      "title": "Error Budget Remaining",
      "type": "gauge",
      "gridPos": {"h": 8, "w": 6, "x": 0, "y": 0}
    },
    {
      "title": "Request Rate",
      "type": "timeseries",
      "gridPos": {"h": 8, "w": 12, "x": 6, "y": 0}
    }
  ]
}

Full Dashboard¶

nthlayer generate-dashboard services/api.yaml --full

Includes all technology-specific panels (28+ panels vs 6-12 in overview mode).

Preview Mode¶

nthlayer generate-dashboard services/api.yaml --dry-run | jq .

With Metric Discovery¶

nthlayer generate-dashboard services/api.yaml \
  --prometheus-url http://prometheus:9090

When Prometheus URL is provided, the command validates that dashboard metrics exist.

Dashboard Panels¶

Overview Mode (Default)¶

Panel	Description
Error Budget	Gauge showing remaining error budget
SLO Status	Current SLO compliance percentage
Request Rate	Requests per second over time
Error Rate	5xx errors as percentage
Latency (p50/p95/p99)	Request latency percentiles
Availability	Uptime percentage

Full Mode (`--full`)¶

Adds technology-specific panels:

PostgreSQL: - Active connections - Transaction rate - Replication lag - Cache hit ratio

Redis: - Memory usage - Hit/miss ratio - Connected clients - Evictions

Kubernetes: - Pod status - Resource utilization - Restart count - Network I/O

Technology Templates¶

Technology	Panels
HTTP/API	8 panels
PostgreSQL	12 panels
Redis	10 panels
Kubernetes	10 panels

Output Structure¶

generated/
└── dashboards/
    ├── payment-api.json
    ├── user-service.json
    └── order-service.json

Importing to Grafana¶

Via API¶

# Generate dashboard
nthlayer generate-dashboard services/api.yaml

# Import to Grafana
curl -X POST \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GRAFANA_TOKEN" \
  -d @generated/dashboards/payment-api.json \
  http://grafana:3000/api/dashboards/db

Via Grafana UI¶

Navigate to Dashboards > Import
Upload the generated JSON file
Select your Prometheus data source

CI/CD Integration¶

jobs:
  generate:
    steps:
      - name: Generate Dashboards
        run: |
          for service in services/*.yaml; do
            nthlayer generate-dashboard "$service" --full
          done

      - name: Upload to Grafana
        run: |
          for dashboard in generated/dashboards/*.json; do
            curl -X POST \
              -H "Authorization: Bearer ${{ secrets.GRAFANA_TOKEN }}" \
              -d @"$dashboard" \
              ${{ vars.GRAFANA_URL }}/api/dashboards/db
          done