Skip to content

Latest commit

 

History

History
210 lines (146 loc) · 9.56 KB

KafkaConsumerQuickStart.md

File metadata and controls

210 lines (146 loc) · 9.56 KB

KafkaConsumer Quick Start

Generally speaking, The Modern C++ Kafka API is quite similar with Kafka Java's API

We'd recommend users to cross-reference them, --especially the examples.

Unlike Java's KafkaConsumer, here we introduced two derived classes, --KafkaAutoCommitConsumer and KafkaManualCommitConsumer, --depending on whether users should call commit manually.

KafkaConsumer (enable.auto.commit=true)

  • Automatically commits previously polled offsets on each poll (and the final close) operations.

    • Note, the internal offset commit is asynchronous, and is not guaranteed to succeed. It's supposed to be triggered (within each poll operation) periodically, thus the occasional failure doesn't quite matter.

Example

        // Create configuration object
        kafka::Properties props ({
            {"bootstrap.servers",  brokers},
            {"enable.auto.commit", "true"}
        });

        // Create a consumer instance
        kafka::clients::KafkaConsumer consumer(props);

        // Subscribe to topics
        consumer.subscribe({topic});

        // Read messages from the topic
        std::cout << "% Reading messages from topic: " << topic << std::endl;
        while (true) {
            auto records = consumer.poll(std::chrono::milliseconds(100));
            for (const auto& record: records) {
                // In this example, quit on empty message
                if (record.value().size() == 0) return 0;

                if (!record.error()) {
                    std::cout << "% Got a new message..." << std::endl;
                    std::cout << "    Topic    : " << record.topic() << std::endl;
                    std::cout << "    Partition: " << record.partition() << std::endl;
                    std::cout << "    Offset   : " << record.offset() << std::endl;
                    std::cout << "    Timestamp: " << record.timestamp().toString() << std::endl;
                    std::cout << "    Headers  : " << kafka::toString(record.headers()) << std::endl;
                    std::cout << "    Key   [" << record.key().toString() << "]" << std::endl;
                    std::cout << "    Value [" << record.value().toString() << "]" << std::endl;
                } else {
                    std::cerr << record.toString() << std::endl;
                }
            }
        }
        
        // consumer.close(); // No explicit close is needed, RAII will take care of it

  • bootstrap.servers property is mandatory for a Kafka client.

  • subscribe could take a topic list. It's a block operation, would wait the consumer to get partitions assigned.

  • poll must be called periodically, thus to trigger kinds of callback handling internally. In practice, it could be put in a "while loop".

  • At the end, we could close the consumer explicitly, or just leave it to the destructor.

KafkaConsumer (enable.auto.commit=false)

  • Users must commit the offsets for received records manually.

Example

        // Create configuration object
        kafka::Properties props ({
            {"bootstrap.servers", brokers},
        });

        // Create a consumer instance
        kafka::clients::KafkaConsumer consumer(props);

        // Subscribe to topics
        consumer.subscribe({topic});

        auto lastTimeCommitted = std::chrono::steady_clock::now();

        // Read messages from the topic
        std::cout << "% Reading messages from topic: " << topic << std::endl;
        bool allCommitted = true;
        bool running      = true;
        while (running) {
            auto records = consumer.poll(std::chrono::milliseconds(100));
            for (const auto& record: records) {
                // In this example, quit on empty message
                if (record.value().size() == 0) {
                    running = false;
                    break;
                }

                if (!record.error()) {
                    std::cout << "% Got a new message..." << std::endl;
                    std::cout << "    Topic    : " << record.topic() << std::endl;
                    std::cout << "    Partition: " << record.partition() << std::endl;
                    std::cout << "    Offset   : " << record.offset() << std::endl;
                    std::cout << "    Timestamp: " << record.timestamp().toString() << std::endl;
                    std::cout << "    Headers  : " << kafka::toString(record.headers()) << std::endl;
                    std::cout << "    Key   [" << record.key().toString() << "]" << std::endl;
                    std::cout << "    Value [" << record.value().toString() << "]" << std::endl;

                    allCommitted = false;
                } else {
                    std::cerr << record.toString() << std::endl;
                }
            }

            if (!allCommitted) {
                auto now = std::chrono::steady_clock::now();
                if (now - lastTimeCommitted > std::chrono::seconds(1)) {
                    // Commit offsets for messages polled
                    std::cout << "% syncCommit offsets: " << kafka::utility::getCurrentTime() << std::endl;
                    consumer.commitSync(); // or commitAsync()

                    lastTimeCommitted = now;
                    allCommitted      = true;
                }
            }
        }

        // consumer.close(); // No explicit close is needed, RAII will take care of it

  • The example is quite similar with the KafkaAutoCommitConsumer, with only 1 more line added for manual-commit.

  • commitSync and commitAsync are both available for a KafkaManualConsumer. Normally, use commitSync to guarantee the commitment, or use commitAsync(with OffsetCommitCallback) to get a better performance.

KafkaConsumer with kafka::clients::KafkaClient::EventsPollingOption

While we construct a KafkaConsumer with kafka::clients::KafkaClient::EventsPollingOption::Auto (i.e. the default option), an internal thread would be created for OffsetCommit callbacks handling.

This might not be what you want, since then you have to use 2 different threads to process the messages and handle the OffsetCommit responses.

Here we have another choice, -- using kafka::clients::KafkaClient::EventsPollingOption::Manual, thus the OffsetCommit callbacks would be called within member function pollEvents().

Example

    KafkaConsumer consumer(props, kafka::clients::KafkaClient::EventsPollingOption::Manual);

    consumer.subscribe({"topic1", "topic2"});

    while (true) {
        auto records = consumer.poll(std::chrono::milliseconds(100));
        for (auto& record: records) {
            // Process the message...
            process(record);

            // Here we commit the offset manually
            consumer.commitSync(*record);
        }

        // Here we call the `OffsetCommit` callbacks
        // Note, we can only do this while the consumer was constructed with `EventsPollingOption::Manual`.
        consumer.pollEvents();
    }

Error handling

No exception would be thrown from a consumer's poll operation.

Instead, once an error occurs, the Error would be embedded in the Consumer::ConsumerRecord.

About Error's value()s, there are 2 cases

  1. Success

    • RD_KAFKA_RESP_ERR__NO_ERROR (0), -- got a message successfully

    • RD_KAFKA_RESP_ERR__PARTITION_EOF, -- reached the end of a partition (no message got)

  2. Failure

Frequently Asked Questions

  • What're the available configurations?

  • How to enhance the polling performance?

    ConsumerConfig::QUEUED_MIN_MESSAGES determines how frequently the consumer would send the FetchRequest towards brokers. The default configuration (i.e, 100000) might not be good enough for small (less than 1KB) messages, and suggest using a larger value (e.g, 1000000) for it.

  • How many threads would be created by a KafkaConsumer?

    Excluding the user's main thread, if enable.auto.commit is false, the KafkaConsumer would start another (N + 2) threads in the background; otherwise, the KafkaConsumer would start (N + 3) background threads. (N means the number of BOOTSTRAP_SERVERS)

    1. Each broker (in the list of BOOTSTRAP_SERVERS) would take a seperate thread to transmit messages towards a kafka cluster server.

    2. Another 3 threads will handle internal operations, consumer group operations, and kinds of timers, etc.

    3. To enable the auto commit, one more thread would be create, which keeps polling/processing the offset-commit callback event.

    E.g, if a KafkaConsumer was created with property of BOOTSTRAP_SERVERS=127.0.0.1:8888,127.0.0.1:8889,127.0.0.1:8890, it would take 6 threads in total (including the main thread).

  • Which one of these threads will handle the callbacks?

    There are 2 kinds of callbacks for a KafkaConsumer,

    1. RebalanceCallback will be triggered internally by the user's thread, -- within the poll function.

    2. If enable.auto.commit=true, the OffsetCommitCallback will be triggered by the user's poll thread; otherwise, it would be triggered by a background thread.