IX25 Performance-Optimierung

Optimale JVM-Flags für IX25 (50-100 Nutzer):

# Heap Size (4-8GB für 50-100 Nutzer)
-Xms4G
-Xmx8G

# G1GC aktivieren
-XX:+UseG1GC

# GC-Pause-Zeit (max 200ms)
-XX:MaxGCPauseMillis=200

# G1 Heap Region Size (16MB optimal für Intrexx)
-XX:G1HeapRegionSize=16M

# GC bei 45% Heap-Belegung starten (default: 65%)
-XX:InitiatingHeapOccupancyPercent=45

# Parallel Reference Processing (schnellere GC)
-XX:+ParallelRefProcEnabled

# System.gc() Calls ignorieren (Apps sollten NICHT manuell GC triggern)
-XX:+DisableExplicitGC

# GC-Logging (für Monitoring)
-XX:+PrintGCDetails
-Xloggc:/var/log/intrexx/gc.log
-XX:+UseGCLogFileRotation
-XX:NumberOfGCLogFiles=5
-XX:GCLogFileSize=10M

Kubernetes HPA Config (deployment.yaml):

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: intrexx-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: intrexx-deployment
  minReplicas: 2          # Minimum 2 Pods (High Availability)
  maxReplicas: 10         # Maximum 10 Pods (bei hoher Last)
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70   # Skaliere bei 70% CPU
  - type: Resource
    resource:
      name: memory
      target:
        type: Utilization
        averageUtilization: 80   # Skaliere bei 80% RAM

resources:
  requests:
    memory: "4Gi"      # Minimum 4GB RAM pro Pod
    cpu: "2"           # Minimum 2 CPU-Cores pro Pod
  limits:
    memory: "8Gi"      # Maximum 8GB RAM pro Pod
    cpu: "4"           # Maximum 4 CPU-Cores pro Pod

livenessProbe:
  httpGet:
    path: /health
    port: 8080
  initialDelaySeconds: 120   # Warte 2 Minuten nach Start
  periodSeconds: 30          # Prüfe alle 30 Sekunden
  timeoutSeconds: 5
  failureThreshold: 3        # 3 Fehler → Pod neu starten

readinessProbe:
  httpGet:
    path: /ready
    port: 8080
  initialDelaySeconds: 60
  periodSeconds: 10
  timeoutSeconds: 3
  failureThreshold: 2        # 2 Fehler → Pod aus Load Balancer nehmen

-- Index auf häufig abgefragte Spalten
CREATE INDEX idx_customers_email ON customers(email);
CREATE INDEX idx_orders_customer_id ON orders(customer_id);
CREATE INDEX idx_orders_created_at ON orders(created_at);

-- Composite Index für Multi-Column-Queries
CREATE INDEX idx_orders_customer_status ON orders(customer_id, status);

-- Index für LIKE-Queries (mit GIN-Index für Volltextsuche)
CREATE INDEX idx_customers_name_gin ON customers USING GIN (to_tsvector('german', name));

❌ Langsam (180ms, Full Table Scan):

SELECT * FROM orders
WHERE customer_id IN (SELECT id FROM customers WHERE email LIKE '%@test.com');

✅ Schnell (60ms, Index-Scan):

SELECT o.* FROM orders o
INNER JOIN customers c ON o.customer_id = c.id
WHERE c.email LIKE '%@test.com';

# Intrexx Portal Manager → Database → Connection Pool
minimumIdle: 10              # Minimum 10 Connections im Pool
maximumPoolSize: 50          # Maximum 50 Connections
connectionTimeout: 20000     # 20 Sekunden Timeout
idleTimeout: 300000          # 5 Minuten Idle-Time
maxLifetime: 1800000         # 30 Minuten Max-Lifetime

docker run -d --name redis \
  -p 6379:6379 \
  -v /data/redis:/data \
  redis:7-alpine \
  redis-server --appendonly yes --maxmemory 2gb --maxmemory-policy allkeys-lru

# Intrexx Portal Manager → System → Cache → Redis
redis.host: localhost
redis.port: 6379
redis.database: 0
redis.timeout: 2000ms
redis.pool.maxActive: 50
redis.pool.maxIdle: 10

# portal.wcf → Session-Management
session.store: redis           # Sessions in Redis statt Database
session.ttl: 3600              # 1 Stunde Session-TTL
session.cookieName: IXSESSION

// Groovy Script: Redis Query-Cache
import redis.clients.jedis.Jedis

def jedis = new Jedis("localhost", 6379)
def cacheKey = "query:customers:active"

// Cache-Check
def cachedResult = jedis.get(cacheKey)
if (cachedResult) {
    return JSON.parse(cachedResult)  // Cache Hit (2ms)
}

// Cache Miss → Database-Query
def result = g_dbAccess.executeQuery("SELECT * FROM customers WHERE active = true")

// Result in Cache speichern (TTL: 5 Minuten)
jedis.setex(cacheKey, 300, JSON.stringify(result))

return result