Cloud Posse
CloudPosse opinion about CDKTF #98
-
|
Hi team, Just as the title says - what is your opinion and\or position on CDKTF? Is it something you are planning on using? Is this something that could be used for official components? |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
|
We've not yet developed a strategy around CDKTF. I can imagine this playing a part at the application layer of the infrastructure (in the "4 Layers of Infrastructure" model). We are starting to extend the component types supported natively by Atmos, having recently released native support for Packer. There are a couple other native component types we're planning as well, including support for CloudFormation and possibly Ansible. |
Beta Was this translation helpful? Give feedback.
-
|
Here's my $0.02. I haven’t used CDKTF directly, but I have worked with its close relative, the AWS CDK, in Java and TypeScript. At first, using a full programming language for infrastructure as code was appealing, and we built a lot with it. However, over time we hit a conceptual wall - the flexibility of the programming language didn’t translate well into the generated Terraform code. For example, even simple refactoring often led to unintended resource replacements. In the end, we abandoned it and went back to Terraform. Obviously, your mileage may vary. |
Beta Was this translation helpful? Give feedback.
We've not yet developed a strategy around CDKTF. I can imagine this playing a part at the application layer of the infrastructure (in the "4 Layers of Infrastructure" model). We are starting to extend the component types supported natively by Atmos, having recently released native support for Packer. There are a couple other native component types we're planning as well, including support for CloudFormation and possibly Ansible.