Chanwit Kaewkasi

@ekristen that's cool. Maybe we would try crafting something and K8s will tell us the real limitation it has.

> or you can just use the k8s client to find and assemble. This would be a good candidate for a `tfctl` command.

Now with the custom backend merged by #288, we can use, for example, S3 to store tfstate. Chunk-based tfplan can now be prioritized.

1 MiB is equal to 1048576 bytes, so we can use 1000000 bytes as the limitation.

A CLI for this would be nice to have

It seems to be a must-fix feature. Let me take a look at it.

Thank you @caspervk This is an excellent addition to TF inputs.

This is an incredible approach. I didn't think about that. We would bundle multiple versions, say 0.1x, 1.1.x, and 1.2.x in the image and make it switchable via a soft...

Closing this for now. We'll support multiple TF binaries shortly.

Hi Vincent, We have a plan but do not support other backends yet. The `backendConfig.disable: true` is designed for Terraform Cloud integration and that's not what you wanted for S3....