1.项目搭建
2.关键代码与配置
3.性能调优
注意,本项目基于spring boot 1,如果是spring boot 2有可能会报错.相应的包需要更新
1.项目搭建
kafka版本:kafka_2.11-1.0.0
jar包版本:1.1.7.REALEASE
<dependency>
<groupId>org.springframework.kafka</groupId>
<artifactId>spring-kafka</artifactId>
<version>1.1..RELEASE</version>
</dependency>
只需要在spring boot工程中加入改jar即可
2.关键代码与配置
实现生产者消费者需要实现几个关键bean
类 KafkaProducerConfig:
import org.apache.kafka.clients.producer.ProducerConfig;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.kafka.annotation.EnableKafka;
import org.springframework.kafka.core.DefaultKafkaProducerFactory;
import org.springframework.kafka.core.KafkaTemplate;
import org.springframework.kafka.core.ProducerFactory; import java.util.HashMap;
import java.util.Map; @Configuration
@EnableKafka
public class KafkaProducerConfig { @Bean("kafkaTemplate")
public KafkaTemplate<String, String> kafkaTemplate() {
KafkaTemplate<String, String> kafkaTemplate = new KafkaTemplate<String, String>(producerFactory());
return kafkaTemplate;
} @Value("${spring.kafka.bootstrap-servers}")
private String kafkaServers; @Value("${spring.kafka.producer.retries}")
private String retry; @Value("${spring.kafka.producer.batch-size}")
private String batch; @Value("${spring.kafka.producer.buffer-memory}")
private String mem; @Value("${spring.kafka.producer.key-serializer}")
private String keySerializer; @Value("${spring.kafka.producer.value-serializer}")
private String valueSerializer; public ProducerFactory<String, String> producerFactory() {
Map<String, Object> properties = new HashMap<String, Object>();
properties.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG,kafkaServers);
properties.put(ProducerConfig.RETRIES_CONFIG, retry);
properties.put(ProducerConfig.BATCH_SIZE_CONFIG, batch);
properties.put(ProducerConfig.LINGER_MS_CONFIG, 1);
properties.put(ProducerConfig.BUFFER_MEMORY_CONFIG, mem);
properties.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, keySerializer);
properties.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, valueSerializer);
return new DefaultKafkaProducerFactory<String, String>(properties);
}
}
几个关键配置:
ProducerConfig.BOOTSTRAP_SERVERS_CONFIG //kafka地址
ProducerConfig.BATCH_SIZE_CONFIG //批量发送配置,单位字节 当多个数据同时发往一个分区时,将被批量控制,减少对服务端的请求
ProducerConfig.BUFFER_MEMORY_CONFIG //生产者缓存,单位字节 生产者对发送数据的缓存总数
现在就构造出了kafkaTemplate对象,可以用他发送消息
kafkaTemplate.send(topic, 0, gson.toJson(Object));
send可以只传三个参数:topic,分区,数据 消费者代码和配置:
类 KafkaConsumerBatchConfig
package com.newland.dc.kafka.kafka; import org.apache.kafka.clients.consumer.ConsumerConfig;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.kafka.annotation.EnableKafka;
import org.springframework.kafka.config.ConcurrentKafkaListenerContainerFactory;
import org.springframework.kafka.config.KafkaListenerContainerFactory;
import org.springframework.kafka.core.ConsumerFactory;
import org.springframework.kafka.core.DefaultKafkaConsumerFactory;
import org.springframework.kafka.listener.AbstractMessageListenerContainer;
import org.springframework.kafka.listener.ConcurrentMessageListenerContainer; import java.util.HashMap;
import java.util.Map; @Configuration
@EnableKafka
public class KafkaConsumerBatchConfig { @Value("${spring.kafka.bootstrap-servers}")
private String servers; @Value("${spring.kafka.consumer.enable-auto-commit}")
private boolean auto; @Value("${spring.kafka.consumer.auto-commit-interval}")
private int interval; @Value("${spring.kafka.consumer.group-id}")
private String group; @Value("${spring.kafka.consumer.auto-offset-reset}")
private String reset; @Value("${spring.kafka.consumer.key-deserializer}")
private String keyDeserializer; @Value("${spring.kafka.consumer.value-deserializer}")
private String valueDeserializer; @Value("${spring.kafka.consumer.max-poll-records:100}")
private String maxPollRecords; @Value("${spring.kafka.consumer.max-poll-interval:1000000}")
private String maxPollInterval; public ConsumerFactory<String, String> consumerFactory() {
Map<String, Object> properties = new HashMap<String, Object>();
properties.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, servers);//注意这里修改为kafka的具体配置项目，我这里只是为了开发演示方便
properties.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, auto);
properties.put(ConsumerConfig.AUTO_COMMIT_INTERVAL_MS_CONFIG, interval);
properties.put(ConsumerConfig.SESSION_TIMEOUT_MS_CONFIG, "15000");
properties.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, keyDeserializer);
properties.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG, valueDeserializer);
properties.put(ConsumerConfig.GROUP_ID_CONFIG, group);
properties.put(ConsumerConfig.MAX_POLL_RECORDS_CONFIG, maxPollRecords);
properties.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, reset);
properties.put(ConsumerConfig.MAX_POLL_INTERVAL_MS_CONFIG, maxPollInterval);
return new DefaultKafkaConsumerFactory<String, String>(properties);
} @Bean
public KafkaListenerContainerFactory<?> batchFactory() {
ConcurrentKafkaListenerContainerFactory<String, String> factory =
new ConcurrentKafkaListenerContainerFactory<>();
factory.setConsumerFactory(consumerFactory());
factory.setConcurrency(1);
factory.setBatchListener(true);//设置为批量消费，每个批次数量在Kafka配置参数中设置ConsumerConfig.MAX_POLL_RECORDS_CONFIG
factory.getContainerProperties().setAckMode(AbstractMessageListenerContainer.AckMode.MANUAL_IMMEDIATE);//设置提交偏移量的方式
return factory;
} }
关键配置:
ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG //由于此处批量我们用手动提交,所以该配置改为false
ConsumerConfig.MAX_POLL_RECORDS_CONFIG //每次批量消费最大数
factory.setBatchListener(true); //注意把批量消费开启
消费者代码:对话题的每个分区监听,注意containerFactory配置
@Component
public class MyListener { @Autowired
private KafkaReceiverBatch kafkaReceiverBatch;
private final Log log = LogFactory.getLogger(MyListener.class); @KafkaListener(id = "id0",containerFactory = "batchFactory", topicPartitions = { @TopicPartition(topic = "${consumer.log.topic:log.business}", partitions = { "0" }) })
public void listenPartition0(List<ConsumerRecord<?, ?>> records, Acknowledgment ack) {
log.info(LogProperty.LOGCONFIG_DEALID,"partition:0, size " + records.size());
kafkaReceiverBatch.batchConsumer(records,ack);
a1 = printNum("0",a += records.size(),a1);
}
@KafkaListener(id = "id1",containerFactory = "batchFactory", topicPartitions = { @TopicPartition(topic = "${consumer.log.topic:log.business}", partitions = { "1" }) })
public void listenPartition1(List<ConsumerRecord<?, ?>> records, Acknowledgment ack) {
log.info(LogProperty.LOGCONFIG_DEALID,"partition:1, size " + records.size());
kafkaReceiverBatch.batchConsumer(records,ack);
b1 = printNum("1",b += records.size(),b1);
}
@KafkaListener(id = "id2",containerFactory = "batchFactory", topicPartitions = { @TopicPartition(topic = "${consumer.log.topic:log.business}", partitions = { "2" }) })
public void listenPartition2(List<ConsumerRecord<?, ?>> records, Acknowledgment ack) {
log.info(LogProperty.LOGCONFIG_DEALID,"partition:2, size " + records.size());
kafkaReceiverBatch.batchConsumer(records,ack);
c1 = printNum("2",c += records.size(),c1);
} static Integer a = 0,b = 0,c = 0;
static Integer a1 = 0,b1 = 0 ,c1 = 0 ;
private Integer printNum(String threadTag, Integer num, Integer printTimes){
if( num/100000 > printTimes ){
System.out.println("partition:" + threadTag + ",consumer num:" + num);
printTimes ++;
}
return printTimes;
}
}
消费逻辑也贴个例子:
protected void batchConsumer(List<ConsumerRecord<?, ?>> records, Acknowledgment ack){
for (ConsumerRecord<?, ?> record : records) {
try {
Optional<?> kafkaMessage = Optional.ofNullable(record.value());
if (kafkaMessage.isPresent()) {
Object message = kafkaMessage.get();
AllLogBase allLogBase = gson.fromJson(message.toString(), AllLogBase.class);
}
} catch (Exception e) {
e.printStackTrace();
continue;
}
}
ack.acknowledge();//手动提交偏移量
}
3.性能调优
kafka生产和消费要注意几个关键点:
1.kafka生产者异步:
pool.execute(()->{kafkaTemplate.send(topic, 0, gson.toJson(Object));});
比如此处可以改为线程池
2.批量写入,可以更改生产者的批量发送值和缓存值,加大该值将大幅提升性能
3.消费者分区监听,并开启批量消费,提升性能