Concept | API Deployer#

Once an API service has been designed and is ready for deployment to an API node, the deployment is accomplished by using the API services component of the Dataiku Deployer (which we refer to as the API Deployer).

The Deployer#

The Dataiku Deployer is accessible from any project in the Design or Automation node, and it consists of a Project Deployer for batch workloads and an API Deployer for real-time processing.

Usage#

Using the API Deployer, you can:

Define API infrastructures, with each infrastructure pointing to an already-installed API node or to a Kubernetes cluster.
Deploy new API services on one or more of these infrastructures.
Monitor the health and status of your API nodes.
Manage the lifecycle of your APIs from development to production. Given that all API services are versioned in the API Deployer, you can roll back to previous versions of an API service at any time.

Deployer setup#

Important

An administrator will be responsible for configuring the Deployer as needed.

Dataiku Cloud users should refer to the documentation on API Node on Dataiku Cloud.
Self-managed users should consult Setting up the Deployer.

With those resources in mind, we’ll briefly cover some key points about the Deployer here.

There are two modes of installing the Deployer:

Mode	Description
Local Deployer	When your infrastructure has a single Design or Automation node that you will use to create API services and/or projects, the Deployer can be part of this Design or Automation node itself. This way, no additional setup is required. This is the method for Dataiku Cloud users.
Remote (or Standalone) Deployer	When your infrastructure includes multiple Design and/or Automation nodes that you will use to create API services and/or projects, you can install a separate Dataiku node which will act as the centralized Deployer for all your Design and Automation nodes.

Deployment infrastructure#

Each deployment infrastructure can either be:

Infrastructure	Description
Static	A set of pre-deployed API nodes that the API Deployer manages.
Kubernetes	The API Deployer dynamically creates containers in your Kubernetes cluster to run the API node server.

Note

A third option is to deploy to an external platform. Learn more in the reference documentation on Deploying to an external platform.

Deployment stages#

These deployment infrastructures may be organized into lifecycle stages (like development, test, and production) that organize how your various deployments will be shown.

What’s next?#

Now that you’ve had an introduction to the basics of the API Deployer in Dataiku, follow Tutorial | Real-time API deployment for practical experience.