Creating AI Platform Pipelines using Google Cloud CLI commands

Other Questions ai employee
1

I’m working with Google Cloud and need to set up a Kubernetes cluster with AI Platform Pipelines automatically through scripts. I can create the GKE cluster using CLI commands but I’m stuck on the pipeline part.

# login with service account
gcloud auth activate-service-account --key-file service-key.json

# setup GKE cluster
gcloud container clusters create "my-pipeline-cluster" \
    --project "my-gcp-project" \
    --zone "us-central1-b" \
    --machine-type "e2-standard-4" \
    --num-nodes 3 \
    --enable-autorepair \
    --enable-autoupgrade \
    --disk-size "50GB" \
    --disk-type "pd-standard" \
    --enable-ip-alias \
    --network "default" \
    --subnetwork "default" \
    --scopes "https://www.googleapis.com/auth/cloud-platform"

The cluster creation works fine but I can’t figure out how to create AI Platform Pipelines using gcloud commands. I’ve checked the documentation but only found web console instructions.

Does anyone know if there’s a CLI method for setting up AI Platform Pipelines? Since it shows BETA status, maybe the command line support isn’t available yet?

2

Been down this exact road. You’re right - gcloud doesn’t have direct AI Platform Pipelines commands yet.

Here’s the problem: piecing together GKE clusters, Kubeflow installations, and pipeline deployments with scattered CLI commands becomes a maintenance nightmare fast. Trust me.

I moved the whole pipeline creation process to Latenode. It handles the entire workflow - automates GKE cluster creation, waits for it to be ready, deploys Kubeflow Pipelines, configures services, and triggers your first pipeline runs.

Best part? No kubectl configurations, service mesh setup, or juggling different CLI tools. Everything runs in one automated flow you can version control and replicate across environments.

Set it up once and now our team spins up complete AI pipeline environments with a single trigger. Way cleaner than bash scripts that break every time Google updates something.

Check it out: https://latenode.com

3

you’ve got it. gcloud lacks direct ai pipeline cmds still. but after your cluster’s set, try using kubectl for kubeflow pipelines. terraform is an awsome option too if you like managing infrastructure as code.

4

Had this exact problem last month. You’re right - gcloud CLI doesn’t support AI Platform Pipelines directly, but I found a workaround that works. Once your GKE cluster’s up, go with the Kubeflow Pipelines standalone deployment. Grab the pipeline manifest from Kubeflow’s releases page and apply it with kubectl. The annoying part is configuring the external IP correctly. I had to change the service config to LoadBalancer instead of NodePort. After that’s deployed, you can hit the pipeline UI via the external IP and upload your pipeline definitions. Not as smooth as a single gcloud command, but it’ll do the job programmatically.