So I’m a Platform Engineer who is currently working mostly on Dockerfiles, Ansible Playbooks and Kubernetes YAMLs (FUCK HELM AND YAML TEMPLATING).
Wanted to know if it’s worth it to invest in learning Pulumi, and advocating for its use in our company? As far as I’ve found out we can unify all of our IaC codes by using Pulumi and get rid of multiple tooling/languages that we currently use + writing tests for our IaC code hopefully. which we do not as of now.
What is Lemmy’s opinion about Pulumi? Is it a shiny new thing that I’m getting hopelessly hyped about because of our current problems, or is it a legit thing that delivers substantial improvements to our flow?
I was also curious about it a few years ago but I never used it in the end, not even to see what it’s like.
But I’m afraid this community you posted on is pretty much dead so I wouldn’t expect many answers :/
Haven’t used it myself, but every person I know that tried Pulumi ended up going back to Terraform. From what I’m given to understand it is fine for small projects, but runs into problems at scale.
Personally I don’t like the default SaaS/account required model in Pulumi. I have lots of things to dislike about Terraform, but that isn’t one.
Mind you too with either tool you (or your devs) will still deal with Dockerfiles and Kubernetes manifests, you just would use Pulumi or Terraform or whatever to manage them.
Lastly: I have done the jump from Ansible to Terraform myself. If you have a large amount of machines to manage and want to minimize transition pains then don’t just use vanilla Terraform, but instead go for a platform like env0 or Spacelift, or at least use Terragrunt to manage your plans.
We’re completely self hosted. Even our k8s cluster is provisioned by us on the baremetal servers we have rented from a third party vendor. So we cannot rely on tools that are not transparently free, seeing pulumis SaaS model is a great downside that I was not paying attention to before you pointing it out.