I'm curious about the reason behind phasing out the NGINX Ingress Controller in favor of the Gateway API, particularly given the current issues with feature compatibility. Many of the annotations that we're reliant on aren't supported or have compatibility issues with the Gateway yet, which complicates migration. I'd like some clarity on the key technical and strategic reasons for this decision, whether there's a roadmap for supporting the commonly used annotations, how migration will work for setups relying on unavailable features, and if any transitional or backward-compatibility measures are planned. Overall, I want to understand how this transition will be executed without disrupting existing workloads. Also, I know the Ingress resource isn't going away, but I want to focus on why some are eager to shift directly to the Gateway API, believing it's the future despite its current shortcomings.
5 Answers
Just to clarify, while the ingress-nginx controller is no longer being maintained, the Ingress API is still fully functional. You'll just need to adopt a different ingress controller, or you can start looking into Gateway API as a strong alternative for future needs. There's no rush to switch—take your time and evaluate your options.
Can you specify which features are lacking in the Gateway API? Understanding these gaps could help in planning the migration better.
People seem confused about the ending of ingress-nginx. It doesn’t mean the Ingress API is disappearing; it’s just a specific controller that’s going away. Ingress still exists in Kubernetes, so you can switch to a different controller like Traefik while continuing to use Ingress. Research the differences between the API and the controllers to avoid misunderstandings!
I think a lot of folks don’t distinguish between the Ingress API and the actual controllers, which leads to a lot of confusion.
The decision to deprecate the NGINX Ingress Controller really stems from a lack of maintainers. The volunteers who were handling it felt overwhelmed, and despite attempts to find more help, they opted to stop development. This has nothing to do with the Gateway API per se—it's simply about the NGINX project. There are still many other Ingress options available if you're not ready to jump into the Gateway just yet.
Absolutely! It's frustrating when large organizations benefit from volunteer work without contributing back. We really need to rethink how we support these projects.
Yes, it’s a tough situation. I completely agree that monetizing OSS projects could help ensure long-term support.
If you're worried about feature gaps with Gateway API, don’t rush it. Many users might decide to stick with a different ingress controller until the Gateway catches up with the needed features. It’s a valid strategy to avoid disruptions while waiting for full compatibility!
Exactly, the ecosystem has plenty of controllers. Just choose one that fits your needs!