nthlayer apply¶

Generate all reliability configs from a service specification.

Usage¶

nthlayer apply <service.yaml> [options]

Options¶

Option	Description
`--push`	Push dashboard to Grafana after generation
`--output-dir DIR`	Output directory (default: `generated/<service>`)
`--dry-run`	Show what would be generated without writing

Examples¶

Basic Generation¶

nthlayer apply payment-api.yaml

Output:

Generated: generated/payment-api/
├── dashboard.json       # Grafana dashboard
├── alerts.yaml          # Prometheus alert rules
├── slos.yaml            # OpenSLO definitions
└── recording-rules.yaml # Prometheus recording rules

Push to Grafana¶

nthlayer apply payment-api.yaml --push

Requires NTHLAYER_GRAFANA_URL and NTHLAYER_GRAFANA_API_KEY to be configured.

Custom Output Directory¶

nthlayer apply payment-api.yaml --output-dir ./configs/payment-api

What Gets Generated¶

Dashboard (dashboard.json)¶

A Grafana dashboard with:

SLO Metrics Row: Availability, latency, error budget
Service Health Row: Request rate, active connections, saturation
Dependencies Row: Panels for each configured dependency (PostgreSQL, Redis, etc.)

Alerts (alerts.yaml)¶

Prometheus alert rules including:

High error rate alerts
Latency threshold alerts
Error budget burn rate alerts
Dependency-specific alerts (e.g., PostgreSQL replication lag)

SLOs (slos.yaml)¶

OpenSLO-compatible definitions:

apiVersion: openslo/v1
kind: SLO
metadata:
  name: payment-api-availability
spec:
  service: payment-api
  indicator:
    metadata:
      name: availability
    spec:
      ratioMetric:
        good:
          metricSource:
            type: prometheus
            spec:
              query: sum(rate(http_requests_total{service="payment-api",status!~"5.."}[5m]))
        total:
          metricSource:
            type: prometheus
            spec:
              query: sum(rate(http_requests_total{service="payment-api"}[5m]))
  objectives:
    - target: 0.9995
      window: 30d

Recording Rules (recording-rules.yaml)¶

Pre-aggregated metrics for dashboard performance:

groups:
  - name: payment-api-recording
    rules:
      - record: service:http_requests:rate5m
        expr: sum(rate(http_requests_total{service="payment-api"}[5m]))
      - record: service:http_errors:rate5m
        expr: sum(rate(http_requests_total{service="payment-api",status=~"5.."}[5m]))