Ask HN: At what stage should I adopt Infrastructure as code?

[sriprasanna]

I am working in a fast paced startup. I feel like writing terraform scripts is slowing us down and I think doing everything manually will be the fastest way go forward. At the same, I don't want to be in a situation where I will be spending lot of time in importing the resources back to code via Terraform at a later stage. In the company, we are all engineers and we are all experts in building the application, and DevOPS isn't our greatest strength and we don't have any budget to hire an expert. Did any of you face a similar situation? How do you strike a balance? At what stage did you end up adopting it and at what cost?

[brudgers]

How good are your Terraform scripts? Improving their creation and maintenance seems like a direct path to improving throughput quickly...and speed is the thing here because you don't have money to hire DevOps of which the important thing is "you don't have money" because the purpose of your job is getting the company to the point where you have money because not having money is how startups die.

Terraform vs IaC isn't what kills startups. It's the wrong problem to prioritize. Without money, there is no future to be more efficient in. Good luck.

[cbushko]

Doing it now will save time in the future, similar to how writing tests now will save you in the future.

If you are using terraform to deploy to the cloud, there are a lot of standard modules out there that you can leverage that will save you from having to write your own.

[infogulch]

Setting up infrastructure manually is like managing code by slapping it in a local directory with no vcs. (Though maybe at different scales.) For the good and bad in both approaches.