It is designed to handle large volume of data with no data loss where real time data processing is allowed.It has high support and community available on the Internet.

It is an open source which can handle real time data, Due to its scalability and fault tolerance it is the best choice for the large amount of real time data processing. No message will be lost even in the case of software or hardware failure

It is a complex to setup for those who are new for the distributed network. It is dependent on the zoo keeper which is a single point failure.

We used Apache Kafka to move our data integration pipelines from batch to streaming. This has given us a more efficient and cost-effective solution to keep our data warehouse up-to-date in near real-time.

Apache Kafka's stream processing capabilities allowed us to move from batch jobs to streaming data pipelines which greatly improved the freshness of our data warehouse.

It's very challenging to set up, manage, and monitor if you're not using a managed SaaS solution.

The functionality of kafka is just so good, everything works fine and nothing breaks any time. also the distribution is really good

Really hard to deploy it, and maintain it. depends on a lot of things like zoopkeeper.

I have usually worked in teams where kafka was already set up and managed by a dedicated team. It has been pretty easy to integrate new services into the existing topics and start consuming messages, there was no friction to populate the data needed in those new services. Another aspect I liked about kafka is that integrating RBAC access to the topics was not hard at all.

The hardest part of using kafka, for me has been the operational overhead it produces, it is really time consuming and expensive to set up a cluster and start working with it, this makes me usually not choose it as a first option when starting a project because of these reasons and use instead something more lightweight like SNS/SQS or rabbit. Using a managed solution like confluent or kinesis has greatly mitigated this issue although still this requires a monthly cost investment before your workload starts generating revenue.

Apache Kafka increases the real time productivity of my team, I can store the data and then process it just because of Apache Kafka, earlier it was difficult for me handle such large amount of data simultaneously.it is scalable and it allows many aaplications to share information flawlessly.kafka makes our work easy and reliable.

Apache Kafka handles large amount of data at a time, speed at which it sends large number of messages is Seamlessly.it's easy to learn and apply Kafka,it allows producer-consumer message flows.The best thing is architecture which I like the most and it is easy to understand/learn how exactly Kafka works.

Apache Kafka is not that cheap. I recommend Apache Kafka only if you are handling large amount of data.one more thing which is don't like much is Kafka's dependency on zookeeper for setting up the cluster and maintenance if it can be removed then it would help.

Been using Kafka to create topics for to and fro event triggered communications. Have achieved best results with respect to point to point communication. Very happy with the functionality and performance

It speed with which data is processes is very high. It can even adjust to any machine or connection failure. Needs very less hardware and it's capacity can be increased at any time without affecting existing performance

Bigger message data is being compressed when processing which affects the performance. Monitoring the queues and topic function is difficult. When multiple queues are being run at the same time the performance is reduced

Have been using to integrate multiple api's and communicate seamlessly. Very satisfied with the solution the app provides

It has very good durability and it can send in many messages from api to api within milli seconds of time. Doesn't need heavy system setup. User friendly. Messages sent to and fro could be retails for defined period of time if needed.

Though it has various messaging capabilities, kafka has poor debugging and logging tools to monitor an app communication topic which is risky. Sometimes when there are multiple queues its performance is reduced drastically

We use it to relay events from upstream systems and consume them in our streaming pipeline. It is very important in our event driven architecture

Easy to use, fast, reliable and has an easy to use API.

It’s a bit hard to maintain when you have a large kafka cluster with a lot of topics.

We did setup the product on on-premise mode. And intended purpose of replacing TIBCO EMS& JMS messaging layer with Kafka is done without much struggle. Considering a open source software, give a try with POC and implement in production systems. You will never regret.

Apache Kafka is open-sourced event streaming platform. We use this technology as an alternative for TIBCO EMS(Which comes with TIBCO BusniessWorks) to cut licensing price when we migrate from TICBO to Mulesoft. It fits well as a messaging layer to communicated to other mule applications.

Looks like a vintage product but really not. True champ.

I really enjoyed to work with Kafka throughout the years on different projects. It has an extensive API and solid documentation in various programming languages. Some of the concepts are really advanced. Kafka helped us to create data processing pipelines and it was quite easy to maintain them.

Kafka includes some really advanced concepts, so for production uses it required highly experienced Software Engineers.

i use it to process many events what sent by iot application i developed.

Kafka is the best stream processing application. it is highly event based. it process data highly speed.

Apache kafka depends on zookeeper. This makes Apache Kafka management difficult.

It has helped us so much in implementing our solution and making it possible to proceed further, to the next steps.

Good with partitioning and replication of data to avoid fault tolerance and avoiding data loss when Kafka goes down. It has very low latency which is important and found it to be scalable, adding in more consumers to read the data when the data published is huge to balance out avoiding Backpressure.

Whenever we publish data from producers to Kafka, it is stored in a buffer and then it is published to Kafka, so we need to always flush the buffer to keep it empty. When the destination server side cannot accept the requests then we need to manually push the message requests from source again multiple times, which could be automated.

1.It's a great product which is preferred over most other messaging systems such as JMS, TIBCO EMS, ActiveMQ, Rabbit MQ etc. because of it's high scalability, throughput and availability. 2.In our use case, we configured apache kafka in our application to process thousands of messages daily instead of conventional TIBCO EMS messaging system. There was a huge difference in performance and it greatly served our purpose.

1.Highly scalable messaging software which can process trillions of messages with almost no latency. 2. Can easily be integrated with multiple products such as integration tools, can be used in modern day programming languages and support in multiple frameworks with community help and large number of library support. 3. Cloud compatible which offers huge amount of storage across multiple regions with high availability. It has very rich user community which comes to rescue if we are stuck anywhere in it's usage and implementation. 4. Helps build fault tolerant applications easily and almost never lets break anything in message processing.

1. It's bit difficult to set it up locally and management of the cluster becomes a little tough when there are huge number of queues are to configured. 2. It matches only exact topic name and does not support wildcard topic selection. 3. Due to compression of the messages when the size increase, it affects the overall performance and the throughput. 4. Lack of messaging paradigms like point-to- point queues, request/reply operations as compared to other systems such as EMS.

The feature set is just amazing, it enables you to build resilient and fault tolerant applications with ease

It's not very easy to install on locally hosted cloud orchestration frameworks like Kubernetes and OpenShift, even though installation on Docker is as simple as running one command with a simple docker-compose file

Capacidade de autogerenciamento e alta disponibilidade

Implementação é muito complexa especialmente para configuração de múltiplas réplicas

The ease of processing large batches of data in a short time and the ease of being able to do it in python. The integration with multiple data sources is something to mention. Its really easy to manage all the incoming data.

the deployment process is not as simple as I would like. Once you know the process it's easy, but as a novice user it can be costly and the errors you get are not intuitive for troubleshooting deployment.