Deploy a stack using mlstacks

Deploying an entire stack with mlstacks.

MLStacks is a Python package that allows you to quickly spin up MLOps infrastructure using Terraform. It is designed to be used with ZenML, but can be used with any MLOps tool or platform. You can deploy a modular MLOps stack for AWS, GCP or K3D using mlstacks. Each deployment type is designed to offer a great deal of flexibility in configuring the resources while preserving the ease of application through the use of sensible defaults.

Check out the full documentation for the mlstacks package for more information.

When should I deploy something using mlstacks?

To answer this question, here are some pros and cons in comparison to the stack-component deploy method which can help you choose what works best for you!

😍 Pros

• Offers a lot of flexibility in what you deploy.

• Deploying with mlstacks gives you a full MLOps stack as the output. Your components and stack is automatically imported to ZenML. This saves you the effort of manually registering all the components.

😥 Cons

• Currently only supports AWS, GCP, and K3D as providers.

• Most stack deployments are Kubernetes-based which might be too heavy for your needs.

• Not all stack components are supported yet.

Installing the mlstacks extra

To install mlstacks, either run pip install mlstacks or pip install "zenml[mlstacks]" to install it along with ZenML.

MLStacks uses Terraform on the backend to manage infrastructure. You will need to have Terraform installed. Please visit the Terraform docs for installation instructions.

MLStacks also uses Helm to deploy Kubernetes resources. You will need to have Helm installed. Please visit the Helm docs for installation instructions.

Deploying a stack

A simple stack deployment can be done using the following command:

zenml stack deploy -p aws -a -n basic -r eu-north-1 -x bucket_name=my_bucket -o sagemaker

This command deploys a stack on AWS that uses an S3 bucket as an artifact store and Sagemaker as your orchestrator. The stack will be imported into ZenML once the deployment is complete and you can start using it right away!

Supported flavors and component types are as follows:

Component Type	Flavor(s)
Artifact Store	s3, gcp, minio
Container Registry	aws, gcp
Experiment Tracker	mlflow
Orchestrator	kubernetes, kubeflow, tekton, vertex
MLOps Platform	zenml
Model Deployer	seldon
Step Operator	sagemaker, vertex

MLStacks currently only supports deployments using AWS, GCP, and K3D as providers.

Want more details on how this works internally?

The stack recipe CLI interacts with the mlstacks repository to fetch the recipes and stores them locally in the Global Config directory.

This is where you could potentially make any changes you want to the recipe files. You can also use native terraform commands like terraform apply to deploy components but this would require you to pass the variables manually using the -var-file flag to the terraform CLI.

CLI Options for zenml stack deploy

Current required options to be passed in to the zenml stack deploy subcommand are:

• -p or --provider: The cloud provider to deploy the stack on. Currently supported providers are aws, gcp, and k3d.

• -n or --name: The name of the stack to be deployed. This is used to identify the stack in ZenML.

• -r or --region: The region to deploy the stack in.

The remaining options relate to which components you want to deploy.

If you want to pass an mlstacks stack specification file into the CLI to use for deployment, you can do so with the -f option. Similarly, if you wish to see more of the Terraform logging, prompts and output, you can pass the -d flag to turn on debug-mode.

Any extra configuration for specific components (as noted in the individual component deployment documentation) can be passed in with the -x option. This option can be used multiple times to pass in multiple configurations.

Interactive stack deployment

If you would like to be guided through the deployment process, you can use the zenml stack deploy command with the --interactive flag. You will still need to provide the provider, name and region options as described above but for the rest, you will get prompts in the CLI as to which components you would like to deploy. For example, using GCP as the provider you might type:

zenml stack deploy -p gcp -n my_new_stack -r us-east1 --interactive

Displaying Terraform outputs for stacks deployed with mlstacks

If you want to view any of the Terraform-generated outputs for a stack deployed with mlstacks, you can do so with the following command:

zenml stack describe -o <STACK_NAME>

This will print any available outputs to the console if you have deployed a stack with mlstacks via ZenML.

Deleting / destroying resources

🗑️ Once you're done running your pipelines, there's only a single command you need to execute that will take care of cleaning up all the resources that you had created on your cloud.

zenml stack destroy <STACK_NAME>

This will offer you the option to delete the underlying stack specifications and state files as well. You can also choose to delete the stack from your ZenML server.