#pulumi (2020-08)
2020-08-28

@Darren Cunningham has joined the channel

we drank the AWS Kool-Aid and don’t have any aspirations of going multi-cloud so we’re doing all of our Cloud IAC w/CDK. However, I am interested in using something like Pulumi or Terraform to IAC our services: GitHub and Okta in particular.
I’m leaning towards Pulumi pretty much solely based on the fact that we can use TypeScript rather than having to learn a proprietary “configuration language”.
Anybody think I’m going to regret that?

Hey man — I’d read through this thread: https://sweetops.slack.com/archives/CQCDCLA1M/p1591892639362700
Really useful discussion regarding this topic from people who have tried out both.
If you’re on the CDK path though… maybe sticking with dynamic IaC tooling would be your companies best bet.
Hey Amazing people.
Any chance Cloud Posse is going to explore Pulumi as a option beyond Terraform.
One argument I keep making often is it might take some time for Pulumi to catch with Community Modules. And I often take Cloud Posse’s high quality Terraform modules to back my argument.
I am currently exploring Pulumi, and would like to hear what you guys have to say.
Is it too early. Would you personally consider Pulumi?
Also is it a big ask for you guys to provision a #pulumi channel to encourage discussion around it?
Thanks.
Pleasant day!

I will say that “learning” HCL / Terraform is not a big on ramp. It’s a friendlier JSON with some higher order abstractions, logic, and functions.

Then again, myself and I’d say generally this community is biased towards Terraform for sure.

awesome, thanks for digging that up!

Honestly @Matt Gowie you’re not wrong about the friendliness. HCL2 is extremely straight forward. It gets a bit complex when you start introducing for_each
or addressing index values, but nothing much beyond “learn the syntax” and enjoy.

That said, I do plan to take another look at Pulumi now that it has a few more months of improvements since I last took a whirl.

I agree that HCL isn’t large lift, but it’s still something. With CDK being TypeScript I’ve had a few engineers be able to submit PRs with minimal AWS specific knowledge which has been great.

Have you checked out cdktf
?


Infrastructure as Code (IaC) is a fundamental component of modern DevOps practices because it enables you to deploy any version of your application infrastructure at will, and facilitates the full lifecycle management of all the resources required to run and monitor your application. Organizations who have adopted DevOps practices often deploy hundreds or even thousands […]

no, I missed this – thank you!