Why 2 HAProxy k8s ingress controller projects?

[mecampbellsoup]

Hi guys -

We are looking at different API gateway solutions for a project we are embarking on.

We wanted to see what HAProxy-as-k8s-ingress options were out there, and immediately you see that in addition to this project there is another:

https://github.com/haproxy-ingress/charts

I asked the other project for a little context on why such a split existed in the community (as compared to e.g. the NGINX/OpenResty gateway people which have all coalesced around Kong), but I was curious to see your perspective on it.

• Are there important differences that potential new users like us should be aware of starting out?
• Is documentation significantly better/worse?
• etc.

Thanks for your help in evaluating our options here.

[daniel-corbett]

Hello 👋, HAProxy Technologies made significant contributions to the jcmoraisjr/haproxy-ingress project, including the use of the Runtime API and DNS for Service Discovery. We saw a demand from our users for an Ingress Controller that we were authoritative on and could address extreme performance requirements. For example, one user was having scaling issues when using the jcmoraisjr/haproxy-ingress project when dealing with thousands of ingress resources, which we had solved due to different architecture decisions within our own implementation. We also need to be able to cater to the needs of our enterprise customers and offer an Ingress Controller that could be extended to support HAProxy Enterprise native modules and enterprise specific features without putting the burden on the community contributors and maintainers. We believe there's room for both options to exist harmoniously and are really happy that multiple solutions exist as it shows the power of open source (and the power of HAProxy 😊 ) and having a diversity of projects is only to the benefit of the overall community.

[kowshikn]

@daniel-corbett Hello, can you share more details about the scaling issue. I am in particular interested to know about any benchmark or production systems that have used haproxytech/kubernetes-ingress with multiple ingress resources. I have a usecase where we will be running multiple ingress resources(about thousand) with high throughput. Would be great if you can share more details about it.

[bluebrown]

But what is the actual difference?

[mecampbellsoup]

But what is the actual difference?

Read the maintainer's response here.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.