Apache Kafka Reviews

The most valuable feature is the documentation, which is good and clear.

An area for improvement would be growth.

I've been using this solution for just over a year.

Kafka works very well.

The initial setup was simple.

I would rate this solution six out of ten.

We had an application stack consisting of Salesforce frontend and a Commander VPN position management system and used Apache Kafka to decouple the microservices. Additionally, we planned to use Kafka for stream processing and to use event sourcing to pull data from legacy systems and reference data to form a compacted topic that the microservices could consume.

The usage of Kafka is a combination of deploying on a personal Kubernetes cluster or using a managed service such as MSK. However, most people who use Kafka are using a managed service provided by Confluent. It can be deployed on the cloud or on-premise.

The most valuable feature of Apache Kafka is its versatility. It can solve many use cases or can be a part of many use cases. Its fundamental value of it is in the real-time processing capability.

You need time-sensitive technology now, particularly in the analytics space. We have looked at using change data capture and Apache Kafka to modernize our analytics capabilities. Additionally, microservices can be used to capture events from legacy systems.

Managing Apache Kafka can be a challenge, but there are solutions. I used the newest release, as it seems they have removed Zookeeper, which should make it easier. Confluent provides a fully managed Kafka platform, in which the cluster does not need to be managed.

If it is a native Apache Kafka, it would have schema registry capabilities. However, this type of functionality is often provided by third-party tools. Additionally, there may be a need for improved manageability and additional tools to manage the cluster, including standard operational metrics and inbuilt management capabilities.

I have been using Apache Kafka for approximately three years.

The solution is highly resilient.

I rate the stability of Apache Kafka a nine out of ten.

Apache Kafka is scalable.

I rate the scalability of Apache Kafka a nine out of ten.

The support from Apache Kafka is good.

The initial setup of Apache Kafka is easy to set up a cluster.  I did the initial setup on my laptop and it is straightforward. I used the Confluent version, but even if you want to run native capabilities it's straightforward to do the implementation.

The recent proof of concept was done on behalf of a client by a system integrator. Similarly, the previous one was mainly done in-house and it utilized Confluent, Apache Kafka, and MSK. The process involved setting up pre-built capabilities.

The price of the solution is low.

I rate the price of Apache Kafka a nine out of ten.

I rate Apache Kafka a nine out of ten.

We are in the early stages of testing this solution in our lab as a demo. It is in development and we are not in production at this point.

We are using this solution to relay events when they happen to multiple receivers at once to allow better functionality.

Apache Kafka has helped our client's online restaurant company by allowing them to take any orders and send the notifications with some other details, such as logic commands, to the different microservices.

The valuable features are the group community and support.

The solution can improve by having automation for developers. We have done many manual calculations and it has been difficult but if it was automated it would be much better.

I have been using this solution for approximately three months.

The solution's scalability is important for our ability to have more throughput from multiple receivers. If we need more throughput it can deliver.

We did use other solutions previously but this solution makes things a lot easier.

The installation is fairly easy. Additionally, there is a cloud-based version available if a use case requires it.

We did the implementation ourselves.

The solution is free, it is open-source.

There is a lot of configuration involved in this solution. We have found many configurations that have helped us but it would be beneficial if there was automation. 

I rate Apache Kafka a five out of ten.

Apache Kafka is used for connecting components between each other in the same application. The use is quite limited, but I was curious about its filtering capability of it.

We implemented the notification system between our components, and we found that Apache Kafka performs well in scalability. It has improved our organization because of the scalability and the comfort of a fail-safe or disaster recovery it provides.

Apache Kafka can improve by making the documentation more user-friendly. It would be beneficial if we could explain to customers in more detail how the solution operates but the documentation get highly technical quickly. For example, if they had a simple page where we can show the customers how it works without the need for the customer to have a computer science background.

I have been using Apache Kafka for approximately two years.

Apache Kafka is scalable. It is easy to add brokers.

We have approximately 30 people using this solution in my organization. They use the solution daily.

I have only used Apache Kafka.

The initial setup of Apache Kafka took some time but after it was easy.

I rate the initial setup of Apache Kafka a three out of five.

We set up the solution in-house.

This is an open-source solution and is free to use.

We have not used the solution in production. We do not have a lot of data at the moment.

I would recommend this solution to others.

I rate Apache Kafka an eight out of ten.

We use Apache Kafka to ingest a lot of data in real time that Apache Spark processes, and the result is used for a tech decision in real time – in the IT environment, infrastructure environment, and IOT environment, like for a  manufacturing plant.

This is an open-source framework. We also sell professional services on this solution and specifically create a business application for customers. 

The application is called Sherlogic. We have two kinds of customers. We have end-user customers that use the Sherlogic solution, and maybe customers don't know that there is Spark and Kafka in Sherlogic. But we have another kind of customer that uses professional services by Xautomata to create tailor-made applications in analytics and the automation process.

We use Apache Kafka for our digital cloud.

To store a large set of analytical data we are using SQL repository. This type of repository works very well but we need specific and high maintenance. The user experience is friendly.

We are looking for alternative solutions, we tried with noSQL solutions and Confluent specific features but the results were not satisfactory both in terms of performance and usability.

We are working on automated SQL repository management and maintenance tools in order to increase the democratization of our platform.

We've been using this solution for a year and a half.

The stability is very nice. We currently manage 50 million events daily.

It's scalable.

Support is good. It's typical for an open source application. You can have all the information in a public portal. If you want specific consulting, there is a company that promotes this consulting worldwide called Conduent. Their consulting is quick and they have a lot of know-how.

It's very complex, like Spark. 

Deployment took 50 minutes for all the Kubernetes ports, Spark, Kafka, and other components based on Sherlogic. In 30 minutes, we created an environment using this program to make installation easier.

Deployment was done in-house, but we're starting a collaboration with another company and we introduced this company to running this solution. Specifically, we started a collaboration with AWS to promote our platform in a Western marketplace. In this way, it's very easy to use our solution because it is a part of an AWS service, certificated by an engineer.

The return on investment has been having people dedicated to this solution because it's open source so it hasn't been necessary to invest in licensing or pay a fee. So, internal know-how has been the ROI.

It's a bit cheaper compared to other Q applications.

I would rate this solution 7 out of 10.

I would recommend this solution because the queue manager is very fast and stable.

One of our clients needed to take events out of SAP to stream them through Apache Kafka while applying data enrichment before reaching the consumers.

The solution can handle more speed and has horizontal scalability for both messaging, but more specifically stream processing and data enrichment. By using this solution it can reduce the number of components required in the tech stack. For example, we were taking data events out of SAP and sending them to consumers without having to go through multiple processors that were outside of the KAFKA space. Additionally, we are using Kafka from GoldenGate to propagate database updates in real-time.

The most valuable features are the stream API, consumer groups, and the way that the scaling takes place. 

I would like to see real-time event-based consumption of messages rather than the traditional way through a loop. The traditional messaging system works by listing and looping with a small wait to check to see what the messages are. A push system is where you have something that is ready to receive a message and when the message comes in and hits the partition, it goes straight to the consumer versus the consumer having to pull. I believe this consumer approach is something they are working on and may come in an upcoming release. However, that is message consumption versus message listening.

Confluent created the KSQL language, but they gave it to the open-source community. I would like to see KSQL be able to be used on raw data versus structured and semi-structured data.

I have been using this solution for approximately one year.

The solution is stable.

I have found the Apache Kafka to be highly scalable

The project we were working on was open-source, we were using Confluent as support and they were great.

Apache Kafka on AWS is a bit complex. There is a third-party company called Confluent and they have the support that makes their installation much easier, especially for the on-premise deployment. You install Apache Kafka alone it can be a little complex compared to other queuing messaging solutions.

The on-premise deployment takes approximately a few days. The cloud or hybrid deployments including all the permissions, typologies, firewalls, and networking configuration can take weeks for all the accessibility issues to be resolved. However, the delay could have been client-related and not necessarily the solution.

We provide the implementation service.

Apache Kafka is free. My clients were using Confluent which provides high-quality support and services, and it was relatively expensive for our client. There was a lot of back and forth on negotiating the price.

Confluent has an offering that has Cloud-Based pricing. There are different packages, prices, and capabilities. The highest level being the most expensive. AWS provides services to their market, for example, to have Kafka running. I do not know what the pricing is and I am fairly confident, Azure and GCP provide similar services.

My advice to others wanting to implement this solution is to start with data streaming projects, not simple messaging projects because while it is very good at general-purpose messaging, it is more suited and geared for when you are using it as a streaming solution.

I rate Apache Kafka an eight out of ten.

I am a solution architect and this is one of the products that I implement for my customers.

Kafka works well when subscribes want to stream data for specific topics.

The most valuable feature is the performance.

Kafka is complex and there is a little bit of a learning curve.

I have been using Apache Kafka for between one and two years.

Resilience-wise, Kafka is very good.

Kafka is a very scalable system. You can have multiple, scalable architectures.

I have not seen any problems with technical support. There is licensed support available, which is not the case with all open-source solutions. Open-source products often have issues when it comes to getting support.

I have customers who were using IBM MQ but they have been switching to open-source.

The initial setup was straightforward for me. However, it is not straightforward for everyone because there are some tricky things to implement. In single-mode it is a little bit easier, but when it is set up as a distributed system then it is more complex because there are a lot of things to be considered.

Kafka is open-source and it is cheaper than any other product.

There is a competing open-source solution called NATS but I see that Apache Kafka is widely used in many places.

Performance-wise, Kafka is better than any of the other products.

This is currently the product that I am recommending to customers. Some customers want an open-source solution.

There are some newer products that are coming on to the market that are even faster than Kafka but this solution is very resilient.

In the long run, I think that open-source will dominate the pace.

I would rate this solution a seven out of ten.

We use the solution for analytics for streaming. We also use it for fraud detection.

The Kafka Streams library gives quite a bit of functionality. The connectors provided by the solution are valuable.

The ability to connect the producers and consumers must be improved. It's still a pain point because a lot of development goes into it.

I have been using the solution for seven to eight years.

For what it does, the tool is very stable. It is a message broker. It receives the messages and holds them for producers and consumers. It's usually everything around Kafka that has stability problems because Kafka does exactly what it's supposed to do.

Scalability is one of the main selling points of the tool. The additional nodes we add give us the additional storage capacity we need. I rate the scalability a ten out of ten. The solution is used across multiple domains in our organization. I use the product daily. It’s a continuously growing platform.

Apache doesn't provide support. There are sites we can go to for information, but there's no support team for Apache. There are companies like Confluent and HPE that provide support for the solution.

We also use Flink and other streaming tools. We use Apache Kafka in addition to other technologies because of the requirement and the business use cases.

It is super easy to set up. I rate the ease of setup a ten out of ten. However, building and administration get quite difficult. It takes three months to make things production-ready.

The deployment was done in-house. We used the tools that we have in our CI/CD pipeline. We needed three people for the deployment. The infrastructure team maintains the tool. The infrastructure team has three to ten members.

We see an ROI on the product. If we don't have a tool to buffer the amount of traffic coming in from high-traffic sites, we cannot use the data. Apache Kafka gives us a resting area where we can push as much information as we want to. It’s picked up by consumers when they need it.

It’s a huge return on investment. Otherwise, we must have a system tied to the producer waiting for the consumer to consume before we can do anything with the rest of the messages. A solution like Kafka provides us with a buffer to consume the data as we choose to.

The price depends on who we are getting the product from. If we buy it from Confluent, we always have to try to negotiate the price. The price is always negotiable.

Overall, I rate the product a six out of ten.