What is our primary use case?
My main use case for Kong Konnect, at least with the clients that I work with, is to create a centralized single pane of glass for all of their control planes, whether it be for their API gateways or their AI gateways, so the SaaS control plane can control and configure and give visibility to all of those data plane nodes.
One of the last clients I worked with was a large automotive enterprise company that uses Kong Konnect to control or house all of their API gateway control planes. Kong Konnect gives you the ability to roll out a federated operating model so that you can have multiple teams; they had about 80 teams, allowing them to not only get self-service to manage their own configurations but also have isolation from anyone else, reducing the blast radius of there being any issues.
What is most valuable?
Some other use cases or features that Kong Konnect gives you that I've seen other customers use include a really good analytics feature that allows you to get information about requests, specifically getting latency information when requests go through the gateway and on which features and plugins are being used. Additionally, having the ability to use an IDP called Kong Identity is a really nice way of doing machine-to-machine identity management rather than bringing in options like Entra. Obviously, I've mentioned API gateways, but the same plays out with AI gateways as well, and having the ability to add the capabilities needed, whether it's linking out to guardrails or being able to use open telemetry to see what's going on. I think the feature set is very rich within Kong.
Kong Konnect reduces the operational overhead of being able to run a control plane, as well as gives you that single pane across large enterprises when you want to work in a federated operating model, which I think is quite a major benefit. It means there's no operational tasks of doing things like updating licenses when you are very federated across an organization, and it just opens up the visibility of what's going through your gateways. I think it's a really great tool, and I've seen a number of enterprise companies using it.
What needs improvement?
The whole area of API and AI management is quite complex, with so many different ways of being able to do things. I think a positive of Kong is that it's so configurable and extendable, but that is also a con because there are so many ways of achieving the same thing that I think often people are either confused or struggle to get going as to how to solve a particular problem. In my mind, a more opinionated deployment of Kong itself or even how to solve certain use cases would be something I would like to see, so a more opinionated use of their products and configuration of their products on how to solve the most common use cases.
That's probably the main thing about needed improvements for Kong Konnect. It's a pretty fully-fledged platform; there obviously are portions of it that could be improved. But I think overall as a platform, it's really good, and I think it can help the majority of organizations looking to manage their throughput of APIs or AI traffic.
The majority of what I've said before probably covers the improvements needed for Kong Konnect. There probably are some API Ops and AI Ops changes that could be improved, which is not necessarily the Kong Konnect product, but some of the toolchains around it. I'm thinking around, especially if it's API Ops, about topics like better GitOps processes and the ways of being able to roll back changes when there are issues. At the moment, a lot of that is either non-existent or is hand-cranked. I would like to see the ability to roll back configurations when there are issues being added to the product.
For how long have I used the solution?
I've been using Kong's product for about seven or eight years, and Kong Konnect has only really been around probably in the last three years, so I've been using Kong Konnect since it first hit the market.
What do I think about the stability of the solution?
Kong Konnect is very stable. Kong Konnect is just the control plane; the data planes or the proxies, which are the gateways, have been around and running for I would say 8 to 10 years, and I've seen them run at major scale in large companies across the whole of Europe. In my opinion, it is very stable.
What do I think about the scalability of the solution?
Kong Konnect can handle scalability. Kong Konnect is the control plane, and there is the ability to add more teams and more control planes based on the number of teams you have onboarding. From that aspect, there's absolutely no problem. When it comes to increased traffic, that's effectively down to how you deploy it. The scalability of the infrastructure, whether it's EKS or ECS as an example in AWS, depends on being able to configure those correctly so that it scales; I have seen some performance metrics that show how well it scales. It can definitely handle growth for any organization.
How are customer service and support?
Customer support for Kong Konnect is really good. Different tiers of support are available depending on how much you want to pay. Overall, the support team is very responsive; it's 24/7 support, so it goes around the world, and you will always have someone in support to talk to. If you aren't getting what you need, there's also the community where they have Kong champions who can answer questions for you as well.
Which solution did I use previously and why did I switch?
I have helped companies migrate from other providers, but I haven't necessarily used those myself. Some of the major reasons behind moving are Kong's ability to follow a proper GitOps model and the simplicity of being able to configure the product as well. Not only that, there is really great support and a great community.
What was our ROI?
I have seen a return on investment from Kong Konnect, though sharing specific numbers is really difficult to do. From my previous client, who was a large enterprise automotive company, what I have seen there is the ability for a smaller team to be able to manage more federated consumers of the services because of how well the federated operating model works with Kong Konnect. Although I can't put a specific number on it, I have seen the scalability of being able to manage 80 to 100-plus teams with a small team of about three or four people.
What's my experience with pricing, setup cost, and licensing?
The processing of the license for Kong Konnect is normally handled by procurement, so in large enterprises, that's not something I've had to deal with, so I can't really comment on that at all. Setup cost is pretty easy to get set up and get going once everything has been licensed and the contract signed, so I wouldn't say that the setup cost is much more compared to using any other product.
Which other solutions did I evaluate?
Before choosing Kong Konnect, I and my clients evaluated other options like the cloud providers' options, such as AWS API Gateway or the Azure API Gateway. Those are often the tools that get compared at least from an API standpoint. When it comes to the AI gateway, there are options like LiteLLM that I've worked with with customers, but I think the features that you get from the full platform of Kong Konnect just outshine something like LiteLLM.
What other advice do I have?
The most simplest case would be a reduction in cost, because you don't have these additional operational tasks that need to happen anymore. Not only does that reduce cost and day-to-day cost, which is time, but it also means allowing teams to self-serve themselves and onboard themselves is a reduction in bottlenecks for platform teams and just works hand-in-hand with the federated operating model.
There is a measurable reduction in onboarding time and certainly a certain percentage of cost savings or resource hours saved after adopting Kong Konnect. I wouldn't be able to put a number on it because I think it will be dependent on the size of the organization, the size of the teams, how federated they are, and how many teams are using the products throughout the organization. It would very much depend on the size of the company for the measurable savings you would have with respect to financial as well as time-based outcomes, but I've definitely seen it improve the self-service capability that a platform team can offer to an organization.
My advice for others looking into using Kong Konnect would be to try it out yourself and just see whether it works and solves what you want within your organization. There is an open-source version, not of Kong Konnect but of the gateway, so you could use that to see whether it works for what you're looking for. I think you can get a two-week trial if you sign up at cloud.konghq.com, so that would be my advice.
I would rate customer support for Kong Konnect with a nine out of ten.
Which deployment model are you using for this solution?
Hybrid Cloud
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
Amazon Web Services (AWS)