crossplane VS tofu-controller

did anyone adopt in production https://crossplane.io ?

- https://github.com/crossplane/crossplane/blob/master/design/one-pager-composition-environment.md

In the overall scheme of things , look at services like backstage.io , crossplane.io and opslevel.com to get ideas. This is not necessarily an endorsement of the services. If all you want is to handle cloud resources and that's it, Terraform can be enough with what ever flavor of web technologies you and your team are comfortable with and can support it along the way. Doesn't take much to create a js based website to collect data from a form, or use other means to collecting data as long as its recorded and transparent for accountability.

XRs are written in a fully declarative manner. And when I am building my XR from underlying managed resources provided by some crossplane provider I need to parametrize resources, use conditionals and create arrays of resuorces The issues of declarativeness in the world of automation are well known- we typically resort to some form of templating and we invent some imperative expressions into that templating language/format. This is currently not very well supported with Crossplane however Crossplane team realizes this issue and they are conteptualizing solution here

I have an idea for a project that would interface with Crossplane. The project has some code that would save tons of time if I could use it directly in my project, but it is located in the internal directory. I can't import the modules directly, but the project is open sourced under an Apache 2.0 license, so the code itself is available for use under that license.

There are always new projects like crossplane that sit on top on architecture systems like terraform, vagrant. The pressure to abstract away any sort of resources is mounting, companies can save a lot by for example by alt hosting S3 endpoints. The train is going the direction not to tie anything to a specific platform implementation if its not a must. Most of the companies I work with use AWS as a hosting provider, but Microsoft for github and related CI matters. As I learned, AWS quality is very dependent on location, eu-central-1 is dead stable for our use cases serving about millions requests a day.

Practically, you'll be replacing stock k8s resources (deployments) with custom ones like Argo Rollouts with Keda autoscaling, so you have to plan the respective Gitops CD pipeline (fluxcd/argocd with some crossplane), as well.

We stumbled upon a project for maintaining Terraform with CRDs that we could deploy with Helm. That project is now called Tofu-Controller - another WeaveWorks project, so it integrated great with our existing Flux setup.

https://github.com/weaveworks/tf-controller/issues/1166#issu...

Weave's TF-Controller, which also has fewer bugs, much better adoption, and it looks like it's actually being developed by someone. But requires a weird argocd <-> flux interop boilerplate. It's a "controller for flux" and not a Kubernetes controller, and I don't really get such ambiguous targeting , but meh...

Wrt tools, I wanted to integrate terraform with Flux thanks to their tf-controller. Conciling the core gitops features with terraform would be great imho.

I'm using Flux instead of Argo which has support for running terraform from a given Git Repo or OCI artifact so essentially I still fall back on Terraform when needed and it's applied via GitOps.

tf-controller which is integrated with Flux GitOps and reconciles Terraform files in a control loop

While searching for alternatives for running Terraform using Kubernetes, I found several controllers and operators, but none that I felt had as much potential as the tf-controller from Weaveworks. We are already using Flux as our GitOps tool, and the tf-controller works by utilizing some of the core functionality from Flux, and has a custom resource for Terraform deployments. The source controller takes care of fetching our modules, the kustomize controllers apply the Terraform resources, and then the controller spin up static pods (called runners) that runs your Terraform commands.

For us, GitOps is a vital part of how we operate, and it is the magic sauce that fuels our platform offering. Not only do we use it for application deployments, but by utilizing the Weaveworks tf-controller, we can create services using Terraform to automate our infrastructure deployments.

Some further research brought me to Weaveworks' TF-Controller which appears to be able to do what I want at least for the initial deployment step. Flux CD (also by Weaveworks) integrates with KEDA now, so it would be great if it could also integrate with KEDA to manage terraform-deployed Azure resources.