Monitoring & Observability

Der Middlelayer ist mit umfassendem Monitoring ausgestattet:

1. Structured Logging (Winston)

Konfiguration:

Log-Level: LOG_LEVEL (default: info)
Format: JSON in Production, farbig in Development
Output: Console + Files (logs/error.log, logs/combined.log)

Verwendung:

import { logger, logQuery, logError } from './monitoring/logger.js';

logger.info('Info message', { context: 'data' });
logQuery('GetProducts', { limit: 10 }, 45);
logError(error, { operation: 'getProducts' });

2. Prometheus Metrics

Endpoints:

GET http://localhost:9090/metrics - Prometheus Metrics
GET http://localhost:9090/health - Health Check

Verfügbare Metriken:

Query Metrics

graphql_queries_total - Anzahl der Queries (Labels: operation, status)
graphql_query_duration_seconds - Query-Dauer (Histogram)
graphql_query_complexity - Query-Komplexität (Gauge)

Cache Metrics

cache_hits_total - Cache Hits (Label: cache_type)
cache_misses_total - Cache Misses (Label: cache_type)

DataService Metrics

dataservice_calls_total - DataService Aufrufe (Labels: method, status)
dataservice_duration_seconds - DataService Dauer (Histogram)

Error Metrics

errors_total - Anzahl der Fehler (Labels: type, operation)

Beispiel Prometheus Query:

# Query Rate
rate(graphql_queries_total[5m])

# Error Rate
rate(errors_total[5m])

# Cache Hit Ratio
rate(cache_hits_total[5m]) / (rate(cache_hits_total[5m]) + rate(cache_misses_total[5m]))

3. Distributed Tracing

Features:

Automatische Trace-ID-Generierung pro Request
Span-Tracking für verschachtelte Operationen
Dauer-Messung für Performance-Analyse

Trace-IDs werden automatisch in Logs und Metrics eingebunden.

Environment Variables

# Logging
LOG_LEVEL=info              # debug, info, warn, error

# Metrics
METRICS_PORT=9090           # Port für Metrics-Endpoint

# Query Complexity
MAX_QUERY_COMPLEXITY=1000   # Max. Query-Komplexität

Integration mit Grafana

Prometheus Scrape Config:

scrape_configs:
  - job_name: 'graphql-middlelayer'
    static_configs:
      - targets: ['localhost:9090']

Grafana Dashboard:

Importiere die Metriken in Grafana
Erstelle Dashboards für:
- Query Performance
- Cache Hit Rates
- Error Rates
- Request Throughput

Beispiel-Dashboard Queries

# Requests pro Sekunde
sum(rate(graphql_queries_total[1m])) by (operation)

# Durchschnittliche Query-Dauer
avg(graphql_query_duration_seconds) by (operation)

# Cache Hit Rate
sum(rate(cache_hits_total[5m])) / (sum(rate(cache_hits_total[5m])) + sum(rate(cache_misses_total[5m])))

# Error Rate
sum(rate(errors_total[5m])) by (type, operation)

2.7 KiB Raw Blame History