Hi <@U04H6UUE78B> - I am going to create Flyte in ...
# flyte-support
b
Hi @average-finland-92144 - I am going to create Flyte in our cloud to MVP Flyte as our new ML production platform. 1. I just started learning terraform, I don't have experience with non-managed clusters. What are the most up-to-date resources that I could follow step-by-step? I found this video

https://www.youtube.com/watch?v=K3xh1X5V0TQ

and https://github.com/unionai-oss/deploy-flyte/tree/main/environments/gcp/flyte-core 2. GCP cloud thank you!
a
Hey Roman, great. Those Terraform modules should help you automating the Flyte deployment. If there's any question during the process please let us know
Also any enhancement to the code and docs for those modules, feel free put it on an Issue
b
👍
@average-finland-92144 'terraform apply' failed several times asking me to install Cloud Resource Manager API, Cloud SQL Admin API, Service Usage API. Cloud Resource Manager API, now it is failing with the followin , screenshot attached: What is the problem? I created a fresh new GCP project to try Flyte.
f
Hi Roman, would it be better to deploy potentially union if you have low experience in terraform? This way you could do a test pretty swiftly
b
I can deploy anything for mybenchmarking tests of Flyte.
f
What does that mean
b
I need to try and eventually to recommend if our company can use Flyte in cloud. We use Slurm, Ray, Kubeflow, and their deployment in cloud is easy.
Out plan was to evaluate Flyte, and if it works compare to Union. Eventually the goal is to see if Union extra features are worth .
f
Got it
b
I would appreciate your advice. How many hours should we expect to spend maintaining Flyte cluster installed on GCP? Perhaps you have statistics of how many hours your clients spend maintaining FLyte clusters in cloud?
f
It’s open source there are many folks that run Flyte. Flyte In our opinion is very resilient, but usecases, scale, integrations matter Cc @brief-window-55364 (Spotify), @cool-lifeguard-49380 (recogni), @adorable-engineer-57446 (Mercedes) may have better answers
b
It's pretty much maintenance free for your everyday case but due to our scale it indeed took us some time to get there, we also have some special cases when it comes to the platform. Make sure you size things correctly (like database) but there's not a size fits all. We did had some surprises in a couple of upgrades but other than that it runs just fine.
Let me know if you have any questions @billowy-glass-15228. Happy to help 🙂
b
Thank you for getting back. Right now I am waiting for a fix for deploying Flyte on GCP from @average-finland-92144. I will give it another try once I hear back from him.
c
Once we had the platform set up and users were onboarded, it was smooth sailing. Work on upgrades or when platform users would like a feature. Also happy to help on GCP @billowy-glass-15228 🙂
But if you would like to maintain it yourself you’ll need someone who is good with K8s, cloud providers, infra as code/terraform etc.
We use Slurm, Ray, Kubeflow, and their deployment in cloud is easy.
Can’t speak for Slurm but Ray and Kubeflow can be installed in an existing K8s cluster with a single
kubectl apply
or
helm install
because all resources that are required are internal to the cluster. Flyte is a bit more complicated than that because it’s a more elaborate but also potent system that makes use of resources outside of the k8s cluster itself like blob storage, managed database, cloud provider IAM permissions, and you’ll need to configure a load balancer and authentication. The terraform module helps with that though.
a
For us Flyte is pretty maintenance free, once everything is up and running. Installing Flyte and extensions like kubeflow, Ray etc. is also quite easy if you know how to handle Helm and Kubernetes in general. Regarding your GCP issues: You need to explicitly enable APIs in order to use products like CloudSQL etc.