Download as PDF, PPTX
val f: java.util.concurrent.Future[RecordMetadata] =
producer.send(new ProducerRecord[String, String]("my-topic", "key", "value"))
f.get // sync
producer.close](https://image.slidesharecdn.com/introductiontokafka-171213143212/85/Introduction-to-Apache-Kafka-33-320.jpg)
consumer.subscribe(Arrays.asList("my-topic"))
while (true) {
val records = consumer.poll(1000) // timeout for long pull
for (record <- records.asScala)
println(record.partition, record.offset, record.key, record.value)
consumer.commitSync
}](https://image.slidesharecdn.com/introductiontokafka-171213143212/85/Introduction-to-Apache-Kafka-34-320.jpg)
val f: java.util.concurrent.Future[RecordMetadata] =
producer.send(new ProducerRecord[String, String]("my-topic", "key", "value"))
f.get // sync
producer.close](https://image.slidesharecdn.com/introductiontokafka-171213143212/75/Introduction-to-Apache-Kafka-33-2048.jpg)
consumer.subscribe(Arrays.asList("my-topic"))
while (true) {
val records = consumer.poll(1000) // timeout for long pull
for (record <- records.asScala)
println(record.partition, record.offset, record.key, record.value)
consumer.commitSync
}](https://image.slidesharecdn.com/introductiontokafka-171213143212/75/Introduction-to-Apache-Kafka-34-2048.jpg)
Uploaded byShiao-An Yuan
Introduction to Apache Kafka
This document provides an introduction to Apache Kafka. It describes Kafka as a distributed messaging system with features like durability, scalability, publish-subscribe capabilities, and ordering. It discusses key Kafka concepts like producers, consumers, topics, partitions and brokers. It also summarizes use cases for Kafka and how to implement producers and consumers in code. Finally, it briefly outlines related tools like Kafka Connect and Kafka Streams that build upon the Kafka platform.
More Related Content
Similar to Introduction to Apache Kafka
![Agentic Systems and Compliance - A brief intro [1.2]](https://cdn.slidesharecdn.com/ss_thumbnails/agenticsystemsandcompliace-1-251018025303-958a42ec-thumbnail.jpg?width=600ounds&width=560&fit=bounds)
![RTP_AR_Basic_Learners' Workbook_KS2 [FOR REPRODUCTION] (1).pdf](https://cdn.slidesharecdn.com/ss_thumbnails/rtparbasiclearnersworkbookks2forreproduction1-251016024943-e51a16ac-thumbnail.jpg?width=600ounds&width=560&fit=bounds)
Introduction to Apache Kafka
- 1. Introduction to ApacheKafka Shiao-An Yuan @sayuan 2017-12-13
- 2. Kafka Producer Kafka Overview Consumer .. 87 6 5 4 3 2 1 0 push pull ZooKeeper ● Developed by LinkedIn ● Written in Scala ● Message queues Topic
- 3. Why We NeedMessage Queues? ● Buffering ● Decoupling
- 4.
- 5.
- 6. Use Cases ● Messaging ●Website Activity Tracking ● Metrics ● Log Aggregation ● Stream Processing ● Event Sourcing ● Commit Log
- 7. Features ● Durability ● Scalability ●Publish-subscribe ● Ordering ● High Availability ● High Throughput ● Delivery Semantics
- 8. Kafka Producer Let’s Add TheseFeatures One by One Consumer .. 8 7 6 5 4 3 2 1 0 push pull Topic
- 9.
- 10. Topic Kafka 9 8 76 5 4 3 2 1 0 Producer Segment Files Consumer
- 11.
- 12. Kafka Broker 1 Partition 0..8 7 6 5 4 3 2 .. .. Broker 2 Partition 1.. 9 8 7 6 5 4 .. .. .. Producer Partitioning Consumer
- 13. Kafka Broker 1 Partition 0..8 7 6 5 4 3 2 .. .. Broker 2 Partition 1.. 9 8 7 6 5 4 .. .. .. Producer Partitioning Consumer Producer Producer
- 14. ConsumerKafka Broker 1 Partition 0..8 7 6 5 4 3 2 .. .. Broker 2 Partition 1.. 9 8 7 6 5 4 .. .. .. Producer Partitioning Producer Producer A1 A2 A3
- 15.
- 16. Kafka Broker 1 Partition 0..8 7 6 5 4 3 2 .. .. Broker 2 Partition 1.. 9 8 7 6 5 4 .. .. .. Producer Consumer Group A Consumer Groups A1 A2 Consumer Group B B1
- 17.
- 18. Consumer Group A Kafka Broker 1 Partition0.. 14 12 10 8 6 4 2 .. .. Broker 2 Partition 1.. 13 11 9 7 5 3 .. .. .. Producer Partitioned by Key A1 A2
- 19.
- 20. Kafka Broker 1 Partition 0 (Leader) ..8 7 6 5 4 3 2 .. .. Partition 1 (Follower).. 9 8 7 6 5 4 .. .. .. Broker 2 Partition 1 (Leader) .. 9 8 7 6 5 4 .. .. .. Partition 0 (Follower) .. 8 7 6 5 4 3 2 .. .. Producer Consumer Replication
- 21. Kafka Broker 1 Partition 0 (Leader) ..8 7 6 5 4 3 2 .. .. Partition 1 (Follower).. 9 8 7 6 5 4 .. .. .. Broker 2 Partition 1 (Leader) .. 9 8 7 6 5 4 .. .. .. Partition 0 (Follower) .. 8 7 6 5 4 3 2 .. .. Producer Consumer Only Produce to / Consume from Leader
- 22. Kafka Broker 1 Partition 0 (Leader) ..8 7 6 5 4 3 2 .. .. Partition 1 (Follower).. 9 8 7 6 5 4 .. .. .. Broker 2 Partition 1 (Leader) .. 9 8 7 6 5 4 .. .. .. Partition 0 (Follower) .. 8 7 6 5 4 3 2 .. .. Producer Consumer Follower Acts Like A Consumer
- 23. High Throughput ● Constanttime suffices ● End-to-end batch compression ● Sequential access ● OS page cache ● Zero-copy
- 24.
- 25.
- 26.
- 28.
- 29. Pulling & Offset Kafka Broker1 Partition 0 (Leader) .. 8 7 6 5 4 3 2 .. .. Broker 2 Partition 1 (Leader) .. 9 8 7 6 5 4 .. .. .. Producer Consumer Group A A1 A2 Consumer Group B B1
- 30. Delivery Semantics ● Atleast once ● At most once ● Exactly once KafkaProducer Consumer
- 31. Delivery Semantics (Produce) ●At least once ○ Retry if no ACK ● At most once ○ Never retry ● Exactly once ○ Idempotence (new in 0.11) KafkaProducer
- 32. Delivery Semantics (Consume) ●At least once ○ Process messages, then commit offsets ● At most once ○ Commit offsets, then process messages ● Exactly once ○ At least once & idempotence Kafka Consumer
- 33. Producer val props =new Properties() props.put("bootstrap.servers", "localhost:9092") props.put("key.serializer", "org.apache.kafka.common.serialization.StringSerializer") props.put("value.serializer", "org.apache.kafka.common.serialization.StringSerializer") val producer = new KafkaProducer[String, String](props) val f: java.util.concurrent.Future[RecordMetadata] = producer.send(new ProducerRecord[String, String]("my-topic", "key", "value")) f.get // sync producer.close
- 34. Consumer val props =new Properties() props.put("bootstrap.servers", "localhost:9092") props.put("group.id", "group1") // consumer group props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer") props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer") val consumer = new KafkaConsumer[String, String](props) consumer.subscribe(Arrays.asList("my-topic")) while (true) { val records = consumer.poll(1000) // timeout for long pull for (record <- records.asScala) println(record.partition, record.offset, record.key, record.value) consumer.commitSync }
- 35. Describe Topic $ bin/kafka-topics.sh--describe --zookeeper "$zookeeper" --topic "test" Topic:test PartitionCount:8 ReplicationFactor:2 Configs:retention.ms=7200000 Topic: test Partition: 0 Leader: 1 Replicas: 1,2 Isr: 1,2 Topic: test Partition: 1 Leader: 2 Replicas: 2,3 Isr: 2,3 Topic: test Partition: 2 Leader: 3 Replicas: 3,4 Isr: 3,4 Topic: test Partition: 3 Leader: 4 Replicas: 4,1 Isr: 1,4 Topic: test Partition: 4 Leader: 1 Replicas: 1,3 Isr: 1,3 Topic: test Partition: 5 Leader: 2 Replicas: 2,4 Isr: 2,4 Topic: test Partition: 6 Leader: 3 Replicas: 3,1 Isr: 1,3 Topic: test Partition: 7 Leader: 4 Replicas: 4,2 Isr: 4,2
- 36. Consumer Offset Checker $bin/kafka-consumer-groups.sh --bootstrap-server "localhost:9092" --describe --group "consumer1" Group Topic Pid Offset logSize Lag consumer1 test 0 60057 80539 20482 consumer1 test 1 47632 66548 18916 consumer1 test 2 11099 30020 18921 consumer1 test 3 45640 65214 19574 consumer1 test 4 60408 61840 1432 consumer1 test 5 76674 96495 19821 consumer1 test 6 63305 82647 19342 consumer1 test 7 06678 25373 18695
- 37.
- 38. Cleanup Policy ● delete ●compact: retain the latest value for each key
- 39.
- 40. Old / NewConsumer API ● Old API ○ --zookeeper localhost:9092 ○ Offsets stored in ZooKeeper ● New API ○ --bootstrap-server localhost:9092 ○ Offsets stored in Kafka’s internal topic
- 41. About Partitions ● Supportincrease the number of partitions only ● Configs (broker & topic) ○ retention.ms ○ retention.bytes ○ segment.bytes
- 42. Adding Brokers ● Reassignedpartitions ○ Load balancing ○ Overhead ○ Throttle ● SiftScience/kafka-assigner ○ Minimized data movement ○ ekoontz/kafka-assigner
- 43.
- 44. Kafka Connect ● Problems ○Schema management ○ Fault tolerance ○ Parallelism ○ Latency ○ Delivery semantics ○ Operations and monitoring ● A framework ● Connectors developed by Confluent ○ S3, Elasticsearch, HDFS, JDBC ● 19 certified connectors developed by vendors ● 67 others connectors listed on official site
- 45. Kafka Stream ● Problems ○Partitioning & scalability ○ Semantics & fault tolerance ○ State ○ Windowing & time ○ Re-processing ● A client library
- 46.
- 47.
- 48.