How Apache Kafka compares to a traditional message queue? In this video, we will go over key features provided by Apache Kafka and see how it compares with traditional message queue systems.
Want to see more videos like this? Let us know what topics you are interested in. Looking for tailored consultations and mentorship sessions? Reach out to us via email at info@relbis.com.
I think that in the "Kafka consumer limit" it should be mentioned that the limit exists within a consumer group, and you can have multiple consumer groups.
At second 4:30, why is that after message 2 fails to be received by consumer 2 and then the process is restarted, the message is sent to consumer 1 instead of consumer 2? I thought that each consumer only reads messages from only 1 partition?
In this case, consumer # 2 stops due to an error and does not recover. After that consumer #1 takes over processing of messages. Sorry for the confusion.
i think the information presented for queue broadcasting is incorrect. Messaging model is either point-to-point or pub-sub. When messaging model is pubsub, we require a message broker inbetween the producer and consumer. The message broker creates copy of the message and it is put into the subscribing MDB(Message Driven Bean). Is it not ?
@KishoreKumar-ur3br, We appreciate your question! 🤝 In this video, we are talking about a traditional message queue. Publish-subscribe messaging pattern uses a different approach. We understand that many modern message brokers support multiple messaging models/patterns. However, this video is not comparing Apache Kafka with a specific implementation of a message broker. The purpose of this video is to compare Apache Kafka with a traditional message queue. You can find the definition of a message queue using the following links: - aws.amazon.com/message-queue/#:~:text=A%20message%20queue%20is%20a,once%2C%20by%20a%20single%20consumer. - en.wikipedia.org/wiki/Message_queue Please feel free to follow up and comment, if something is not clear. 👍
This is all such nonsense. The major queue products support the same functionality as Kafka. It’s entirely a matter of how you choose to configure them.
@@relbis-labs there is no reason to use it. We lived without Kafka for decades, we survived. I've used Kafka in the past, even used it for personal projects. Discovered it was a worthless overcomplicated tool.
A lot of people who use it have a lot of reasons to do so. As in any engineering discipline, we must use the right tool for the job. Still really appreciate you sharing your thoughts and your opinion, even if it does not align with our point of view. Having diverse opinions helps everyone to learn and grow.
