Skip to content

grafana-prometheus-krakend

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

grafana-prometheus-krakend

README.md

Grafana + Prometheus + KrakenD API Gateway

This stack provides a production-ready API Gateway using KrakenD with observability via Prometheus and Grafana.

How it works

flowchart LR
    Client([Client]) -->|:8080| KrakenD[KrakenD Gateway]
    KrakenD --> Backend[Backend Services]
    KrakenD -->|:9091 /metrics| Prometheus[Prometheus :9090]
    Prometheus --> Grafana[Grafana :3000]
Loading
  1. KrakenD runs as a stateless API Gateway
  2. Routes requests to backend services
  3. Exposes metrics at /metrics on port 9091
  4. Prometheus scrapes metrics from KrakenD
  5. Grafana visualizes collected metrics

Architecture Overview

KrakenD (8080)
   |
   |--> Metrics (/metrics:9091)
   |
Prometheus (9090)
   |
Grafana (3000)

Services

  • KrakenD API Gateway
  • Prometheus Metrics Collector
  • Grafana Dashboard

Ports

Service URL
KrakenD API http://localhost:8080
Metrics http://localhost:9091/metrics
Prometheus http://localhost:9090
Grafana http://localhost:3000

KrakenD Configuration (IMPORTANT)

Metrics are exposed using OpenTelemetry Prometheus exporter:

"extra_config": {
  "telemetry/logging": {
    "level": "INFO",
    "prefix": "[KRAKEND]",
    "stdout": true
  },

  "telemetry/opentelemetry": {
    "service_name": "krakend-gateway",
    "exporters": {
      "prometheus": [
        {
          "name": "local_prometheus",
          "port": 9091,
          "process_metrics": true,
          "go_metrics": true
        }
      ]
    }
  }
}

Prometheus Configuration

global:
  scrape_interval: 15s
  evaluation_interval: 15s

scrape_configs:
  - job_name: "prometheus"
    static_configs:
      - targets: ["prometheus:9090"]

  - job_name: "krakend"
    static_configs:
      - targets: ["krakend:9091"]

Docker Compose

version: "3.9"

services:
  krakend:
    image: devopsfaith/krakend:latest
    container_name: krakend
    restart: unless-stopped
    ports:
      - "8080:8080"
      - "9091:9091"
    volumes:
      - ./data/krakend.json:/etc/krakend/krakend.json:ro
    command: ["run", "-c", "/etc/krakend/krakend.json"]
    networks:
      - monitoring

  prometheus:
    image: prom/prometheus:latest
    container_name: prometheus
    restart: always
    ports:
      - "9090:9090"
    volumes:
      - ./prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
      - prometheus_data:/prometheus
    command:
      - "--config.file=/etc/prometheus/prometheus.yml"
      - "--storage.tsdb.path=/prometheus"
    networks:
      - monitoring

  grafana:
    image: grafana/grafana:latest
    container_name: grafana
    restart: always
    ports:
      - "3000:3000"
    environment:
      GF_SECURITY_ADMIN_USER: admin
      GF_SECURITY_ADMIN_PASSWORD: admin
    volumes:
      - grafana_data:/var/lib/grafana
    depends_on:
      - prometheus
    networks:
      - monitoring

volumes:
  prometheus_data:
  grafana_data:

networks:
  monitoring:
    driver: bridge

Key Endpoints

KrakenD API

http://localhost:8080/github/user

Metrics Endpoint

http://localhost:9091/metrics

Validation Steps

1. Check KrakenD API

curl http://localhost:8080/github/user

2. Check Metrics

curl http://localhost:9091/metrics

3. Check Prometheus Targets

http://localhost:9090/targets

Notes

  • /metrics is exposed by KrakenD via OpenTelemetry Prometheus exporter
  • No need for json-exporter anymore
  • Prometheus scrapes directly from KrakenD
  • Ensure port 9091 is exposed in Docker

Recommended Improvements

  • Add Alertmanager for alerting
  • Add Loki for logs
  • Add distributed tracing via OTEL Collector
  • Add Kubernetes Helm deployment

Summary

✔ Direct Prometheus scraping from KrakenD ✔ No JSON exporter required ✔ OpenTelemetry-based metrics pipeline ✔ Simplified architecture