Visão Geral
O Curso Kafka Performance Tuning & Optimization foi projetado para capacitar profissionais na análise, ajuste e otimização de desempenho de clusters Apache Kafka.
Durante o treinamento, os participantes aprenderão a identificar gargalos, configurar parâmetros críticos e aplicar técnicas avançadas para melhorar throughput, reduzir latência e aumentar a eficiência operacional.
Com foco em ambientes de produção e cenários reais, o curso combina teoria e prática para aprimorar a performance de producers, consumers e brokers em escala corporativa.
Conteúdo Programatico
Module 1: Introduction to Kafka Performance
- Understanding performance architecture in Kafka
- Key metrics: throughput, latency, ISR, and partition load
- Common performance bottlenecks and their root causes
- Overview of performance tuning methodology
Module 2: Broker-Level Optimization
- Broker configuration parameters impacting performance
- Tuning replication, request handling, and network threads
- Log segment management and I/O tuning
- Disk and filesystem optimization strategies
Module 3: Producer Performance Tuning
- Producer batching, compression, and acks strategies
- Configuring linger.ms, batch.size, and buffer.memory
- Choosing optimal serializers and compression codecs
- Testing and benchmarking producer throughput
Module 4: Consumer Performance Tuning
- Consumer group coordination and rebalancing
- Fetch size, max.poll settings, and parallelism tuning
- Handling large message loads and backpressure
- Reducing consumer lag and improving processing rate
Module 5: Cluster and Topic Design Optimization
- Topic partitioning strategies for scalability
- Balancing partitions and brokers
- Retention policies and compaction tuning
- Planning hardware resources for optimal performance
Module 6: JVM, OS, and Hardware Optimization
- Kafka JVM tuning (heap, GC, memory management)
- Linux kernel tuning for I/O and networking
- CPU, memory, and disk best practices for Kafka nodes
- Benchmarking hardware and network performance
Module 7: Monitoring and Benchmarking Tools
- Using JMX metrics, Prometheus, and Grafana dashboards
- Understanding key metrics for producers, consumers, and brokers
- Running performance tests with Kafka Performance Tools
- Using open-source tools like Cruise Control and Kafdrop
Module 8: Advanced Optimization and Troubleshooting
- Performance profiling under load
- Detecting and resolving producer/consumer lag
- Handling network congestion and throttling
- Fine-tuning replication and leader election for stability
Module 9: Hands-On Project
Project: Design and tune a Kafka cluster for high throughput and low latency. Perform benchmarking, analyze metrics, and implement configuration adjustments for performance improvement.