upstream-seaweedfs/test/kafka/kafka-client-loadtest/config/loadtest.yaml

# Kafka Client Load Test Configuration

# Test execution settings
test_mode: "comprehensive"  # producer, consumer, comprehensive
duration: "60s"  # Test duration (0 = run indefinitely) - producers will stop at this time, consumers get +120s to drain

# Kafka cluster configuration
kafka:
  bootstrap_servers:
    - "kafka-gateway:9093"
  # Security settings (if needed)
  security_protocol: "PLAINTEXT"  # PLAINTEXT, SSL, SASL_PLAINTEXT, SASL_SSL
  sasl_mechanism: ""  # PLAIN, SCRAM-SHA-256, SCRAM-SHA-512
  sasl_username: ""
  sasl_password: ""

# Schema Registry configuration
schema_registry:
  url: "http://schema-registry:8081"
  auth:
    username: ""
    password: ""

# Producer configuration
producers:
  count: 10  # Number of producer instances
  message_rate: 1000  # Messages per second per producer
  message_size: 1024  # Message size in bytes
  batch_size: 100  # Batch size for batching
  linger_ms: 5  # Time to wait for batching
  compression_type: "snappy"  # none, gzip, snappy, lz4, zstd
  acks: "all"  # 0, 1, all
  retries: 3
  retry_backoff_ms: 100
  request_timeout_ms: 30000
  delivery_timeout_ms: 120000
  
  # Message generation settings
  key_distribution: "random"  # random, sequential, uuid
  value_type: "avro"  # json, avro, protobuf, binary
  schema_format: ""  # AVRO, JSON, PROTOBUF - schema registry format (when schemas enabled)
                     # Leave empty to auto-distribute formats across topics for testing:
                     #   topic-0: AVRO, topic-1: JSON, topic-2: PROTOBUF, topic-3: AVRO, topic-4: JSON
                     # Set to specific format (e.g. "AVRO") to use same format for all topics
  include_timestamp: true
  include_headers: true

# Consumer configuration  
consumers:
  count: 5  # Number of consumer instances
  group_prefix: "loadtest-group"  # Consumer group prefix
  auto_offset_reset: "earliest"  # earliest, latest
  enable_auto_commit: true
  auto_commit_interval_ms: 100  # Reduced from 1000ms to 100ms to minimize duplicate window
  session_timeout_ms: 30000
  heartbeat_interval_ms: 3000
  max_poll_records: 500
  max_poll_interval_ms: 300000
  fetch_min_bytes: 1
  fetch_max_bytes: 52428800  # 50MB
  fetch_max_wait_ms: 100  # 100ms - very fast polling for concurrent fetches and quick drain

# Topic configuration
topics:
  count: 5  # Number of topics to create/use
  prefix: "loadtest-topic"  # Topic name prefix
  partitions: 4  # Partitions per topic (default: 4)
  replication_factor: 1  # Replication factor
  cleanup_policy: "delete"  # delete, compact
  retention_ms: 604800000  # 7 days
  segment_ms: 86400000  # 1 day

# Schema configuration (for Avro/Protobuf tests)
schemas:
  enabled: true
  registry_timeout_ms: 10000
  
  # Test schemas
  user_event:
    type: "avro"
    schema: |
      {
        "type": "record",
        "name": "UserEvent",
        "namespace": "com.seaweedfs.test",
        "fields": [
          {"name": "user_id", "type": "string"},
          {"name": "event_type", "type": "string"},
          {"name": "timestamp", "type": "long"},
          {"name": "properties", "type": {"type": "map", "values": "string"}}
        ]
      }
      
  transaction:
    type: "avro" 
    schema: |
      {
        "type": "record",
        "name": "Transaction", 
        "namespace": "com.seaweedfs.test",
        "fields": [
          {"name": "transaction_id", "type": "string"},
          {"name": "amount", "type": "double"},
          {"name": "currency", "type": "string"},
          {"name": "merchant_id", "type": "string"},
          {"name": "timestamp", "type": "long"}
        ]
      }

# Metrics and monitoring
metrics:
  enabled: true
  collection_interval: "10s"
  prometheus_port: 8080
  
  # What to measure
  track_latency: true
  track_throughput: true
  track_errors: true
  track_consumer_lag: true
  
  # Latency percentiles to track
  latency_percentiles: [50, 90, 95, 99, 99.9]

# Load test scenarios
scenarios:
  # Steady state load test
  steady_load:
    producer_rate: 1000  # messages/sec per producer
    ramp_up_time: "30s"
    steady_duration: "240s" 
    ramp_down_time: "30s"
    
  # Burst load test  
  burst_load:
    base_rate: 500
    burst_rate: 5000
    burst_duration: "10s"
    burst_interval: "60s"
    
  # Gradual ramp test
  ramp_test:
    start_rate: 100
    end_rate: 2000
    ramp_duration: "300s"
    step_duration: "30s"

# Error injection (for resilience testing)
chaos:
  enabled: false
  producer_failure_rate: 0.01  # 1% of producers fail randomly
  consumer_failure_rate: 0.01  # 1% of consumers fail randomly
  network_partition_probability: 0.001  # Network issues
  broker_restart_interval: "0s"  # Restart brokers periodically (0s = disabled)

# Output and reporting
output:
  results_dir: "/test-results"
  export_prometheus: true
  export_csv: true
  export_json: true
  real_time_stats: true
  stats_interval: "30s"
  
# Logging
logging:
  level: "info"  # debug, info, warn, error
  format: "text"  # text, json
  enable_kafka_logs: false  # Enable Kafka client debug logs