What is our primary use case?
Kpow for Apache Kafka is used for persistent messages, payment-type messages, processing log-type messages, consuming data in general from different systems, part of an information infrastructure or an information architecture, consuming information from various sources, then storing that information in a database and processing that information. Processing payment messages is a big application, also for integration and storing the messages persistently, and processing them concurrently.
I have used Kpow for Apache Kafka in partition mode where someone else has set it up with message partitions or topic partitions, and I then consume messages from specific partitions; however, I haven't set it up myself in that mode.
I have not used Kpow for Apache Kafka consumer lag insights; I've seen it and heard about it, but I haven't used that so far. I've only used it where the messages are consumed into some other system that does something with the messages.
What is most valuable?
I see benefits from the product in terms of cost; replacing IBM MQ with Kpow for Apache Kafka is much cheaper. Using Kafka instead of something such as IBM MQ is much cheaper, offering scalability and processing messages in parallel, which Kafka helps manage quite a lot, though you can have issues with duplicate processing. Cost-wise and in terms of simplicity in setup versus something such as IBM MQ, that tends to be a bit more complicated to set up.
The product's adaptability is quite important for my organization as that's the central part of how the information from external systems are processed.
What needs improvement?
I think Kpow for Apache Kafka could improve in the area of monitoring; you can change many parts of it, and it can run out of storage if you don't monitor it, which could cause issues. Having the monitoring perhaps on by default would be good, but besides space issues, I don't see anything else that could be improved.
For how long have I used the solution?
I have been working with Kpow for Apache Kafka for around ten years, on and off, as I worked extensively on it about ten years ago and then again in the meantime.
What do I think about the stability of the solution?
I have not used Kpow for Apache Kafka consumer lag insights; I've seen it and heard about it, but I haven't used that so far. I've only used it where the messages are consumed into some other system that does something with the messages.
What do I think about the scalability of the solution?
I can't really assess the product's ability to identify performance bottlenecks in Kafka clusters because I haven't managed the clusters myself too much. I would think it would be similar to how you scale and manage a Kubernetes cluster, launching more pods or scaling out the cluster if the throughput is not high enough.
How are customer service and support?
My thoughts on technical support involve finding a lot of information more from open-source forums and the internet; I think there is a company that does support, specifically Confluent, as a commercial provider for Kafka that could be used for support, but I haven't worked on any support issues. Most of the issues are what people find online regarding managing Kafka.
I would rate the work of support as typically a six or a seven based on open-source information.
How was the initial setup?
The deployment for Kpow for Apache Kafka can take just minutes for local setup; setting it up in a clustered environment might take a couple of hours, but for anyone knowing how to do that, once you know how to set it up in a Kubernetes cluster or just with a Kafka cluster, then that should take minutes.
One person can do it; it should be quite simple for one person to set up.
What's my experience with pricing, setup cost, and licensing?
I am saying that the cloud version is quite expensive, and there's room for improvement since I've set up a test cluster on my own AWS account, and within the first couple of days, it already accumulated a bill close to $200-$300 with no activity on the cluster. If I didn't stop it that month, I would have gotten a bill close to $1,000 and it wasn't doing much.
I am talking about the public cloud, AWS, which is quite expensive even if you don't run much on the topics or so, and so I shut it down there.
I understand it was not purchased through AWS Marketplace; I just set up via the console, a test cluster, you set up a Kafka test cluster as you would set up an instance, and they bill you for that, which turned out to be quite expensive, as many of the latest services on AWS, especially database services also, RDS or so.
Which other solutions did I evaluate?
I have thought of typical alternatives to Kpow for Apache Kafka, including RabbitMQ or IBM MQ depending on whether an organization is using IBM MQ; ActiveMQ is also a popular alternative, and AWS has Kinesis that can be used for streaming of messages.
What other advice do I have?
I did not face any challenges during deployment with Apache SkyWalking; it was quite straightforward. I haven't worked with that.
I have not worked with telemetry tools, and I haven't used those.
The product's adaptability is quite important for my organization as that's the central part of how the information from external systems are processed.
I would rate this product nine out of ten overall.
Which deployment model are you using for this solution?
On-premises
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
Amazon Web Services (AWS)