我使用 apache camel kafka 作为生成消息的客户端,我观察到 kafka 生产者需要 1 毫秒才能推送一条消息,如果我使用骆驼聚合将消息合并到批处理中,那么推送一条消息需要 100 毫秒。
安装简述
3 kafka 集群 16 核 32GB RAM
示例代码
String endpoint="kafka:test?topic=test&brokers=nodekfa:9092,nodekfb:9092,nodekfc:9092&lingerMs=0&maxInFlightRequest=1&producerBatchSize=65536";
Message message = new Message();
String payload = new ObjectMapper().writeValueAsString(message);
StopWatch stopWatch = new StopWatch();
stopWatch.watch();
for (int i=0;i<size;i++)
{
producerTemplate.sendBody(endpoint,ExchangePattern.InOnly, payload);
}
logger.info("Time taken to push {} message is {}",size,stopWatch.getElasedTime());
骆驼生产者端点
kafka:[topic]?topic=[topic]&brokers=[brokers]&maxInFlightRequest=1
尽管 kafka 文档吹嘘生产者 tps 约为 100,000,但我得到的吞吐量为 1000/s。
如果camel-kafka或kafka本身有任何错误,请告诉我。
生产者配置
acks = 1
batch.size = 65536
bootstrap.servers = [nodekfa:9092, nodekfb:9092, nodekfc:9092]
buffer.memory = 33554432
client.id =
compression.type = none
connections.max.idle.ms = 540000
enable.idempotence = false
interceptor.classes = []
key.serializer = class org.apache.kafka.common.serialization.StringSerializer
linger.ms = 0
max.block.ms = 60000
max.in.flight.requests.per.connection = 1
max.request.size = 1048576
metadata.max.age.ms = 300000
metric.reporters = []
metrics.num.samples = 2
metrics.recording.level = INFO
metrics.sample.window.ms = 30000
partitioner.class = class org.apache.kafka.clients.producer.internals.DefaultPartitioner
receive.buffer.bytes = 65536
reconnect.backoff.max.ms = 1000
reconnect.backoff.ms = 50
request.timeout.ms = 305000
retries = 0
retry.backoff.ms = 100
sasl.client.callback.handler.class = null
sasl.jaas.config = null
sasl.kerberos.kinit.cmd = /usr/bin/kinit
sasl.kerberos.min.time.before.relogin = 60000
sasl.kerberos.service.name = null
sasl.kerberos.ticket.renew.jitter = 0.05
sasl.kerberos.ticket.renew.window.factor = 0.8
sasl.login.callback.handler.class = null
sasl.login.class = null
sasl.login.refresh.buffer.seconds = 300
sasl.login.refresh.min.period.seconds = 60
sasl.login.refresh.window.factor = 0.8
sasl.login.refresh.window.jitter = 0.05
sasl.mechanism = GSSAPI
security.protocol = PLAINTEXT
send.buffer.bytes = 131072
ssl.cipher.suites = null
ssl.enabled.protocols = [TLSv1.2, TLSv1.1, TLSv1]
ssl.endpoint.identification.algorithm = https
ssl.key.password = null
ssl.keymanager.algorithm = SunX509
ssl.keystore.location = null
ssl.keystore.password = null
ssl.keystore.type = JKS
ssl.protocol = TLS
ssl.provider = null
ssl.secure.random.implementation = null
ssl.trustmanager.algorithm = PKIX
ssl.truststore.location = null
ssl.truststore.password = null
ssl.truststore.type = JKS
transaction.timeout.ms = 60000
transactional.id = null
value.serializer = class org.apache.kafka.common.serialization.StringSerializer
测试日志
DEBUG [2019-06-02 17:30:46,781] c.g.p.f.u.AuditEventNotifier: >>> Took 3 millis for the exchange on the route : null
DEBUG [2019-06-02 17:30:46,781] c.g.p.f.u.AuditEventNotifier: >>> Took 3 millis to send to external system : kafka://test?brokers=nodekfa%3A9092%2Cnodekfb%3A9092%2Cnodekfc%3A9092&lingerMs=0&maxInFlightRequest=1&producerBatchSize=65536&topic=test by thead http-nio-8551-exec-6
DEBUG [2019-06-02 17:30:46,783] c.g.p.f.u.AuditEventNotifier: >>> Took 2 millis for the exchange on the route : null
DEBUG [2019-06-02 17:30:46,783] c.g.p.f.u.AuditEventNotifier: >>> Took 2 millis to send to external system : kafka://test?brokers=nodekfa%3A9092%2Cnodekfb%3A9092%2Cnodekfc%3A9092&lingerMs=0&maxInFlightRequest=1&producerBatchSize=65536&topic=test by thead http-nio-8551-exec-6
DEBUG [2019-06-02 17:30:46,784] c.g.p.f.u.AuditEventNotifier: >>> Took 1 millis for the exchange on the route : null
DEBUG [2019-06-02 17:30:46,785] c.g.p.f.u.AuditEventNotifier: >>> Took 2 millis to send to external system : kafka://test?brokers=nodekfa%3A9092%2Cnodekfb%3A9092%2Cnodekfc%3A9092&lingerMs=0&maxInFlightRequest=1&producerBatchSize=65536&topic=test by thead http-nio-8551-exec-6
DEBUG [2019-06-02 17:30:46,786] c.g.p.f.u.AuditEventNotifier: >>> Took 1 millis for the exchange on the route : null
DEBUG [2019-06-02 17:30:46,786] c.g.p.f.u.AuditEventNotifier: >>> Took 1 millis to send to external system : kafka://test?brokers=nodekfa%3A9092%2Cnodekfb%3A9092%2Cnodekfc%3A9092&lingerMs=0&maxInFlightRequest=1&producerBatchSize=65536&topic=test by thead http-nio-8551-exec-6
DEBUG [2019-06-02 17:30:46,788] c.g.p.f.u.AuditEventNotifier: >>> Took 2 millis for the exchange on the route : null
DEBUG [2019-06-02 17:30:46,788] c.g.p.f.u.AuditEventNotifier: >>> Took 2 millis to send to external system : kafka://test?brokers=nodekfa%3A9092%2Cnodekfb%3A9092%2Cnodekfc%3A9092&lingerMs=0&maxInFlightRequest=1&producerBatchSize=65536&topic=test by thead http-nio-8551-exec-6
INFO [2019-06-02 17:30:46,788] c.g.p.f.a.MessageApiController: Time taken to push 5 message is 10ms
显然,消息至少需要 1 毫秒,默认工作池最大大小为 20 ,如果我将压缩编解码器设置为 snappy 这将使性能最差。
让我知道我错过了什么!