Oracle Cloud Infrastructure Documentation

Creating a Deployment

Create a deployment in OCI Generative AI service to deploy an agent container that can serve client requests for an application.

When you create a deployment, you select a container image and image tag. You can also activate the deployment during creation to make it the running version of the application.

This task describes how to create a deployment by using the Oracle Cloud Console.

Resource Usage and Scaling

Deployment usage depends on the resources that run the active deployment. Each running replica uses Compute resources, and any managed OCI services enabled for the application might create other usage.

Application settings control how deployments scale. These settings include the minimum number of replicas, maximum number of replicas, and autoscaling metrics. See Scaling.

A deployment runs as one or more replicas. Each running replica uses the service-defined Compute allocation for the application. When autoscaling adds replicas, Compute usage increases based on the number of running replicas.

For example, if an active deployment scales from one replica to three replicas, it uses three replicas of Compute for the time that those replicas are running.

Tip

To estimate Compute usage, use the following formula:

average running replicas x runtime hours x Compute allocation per replica

The average number of running replicas depends on the application minimum replicas, maximum replicas, traffic volume, and autoscaling settings.

To limit scale-out, set a smaller maximum replica count. To disable autoscaling, set the minimum replicas and maximum replicas to the same value.

Managed storage, cache, database, logging, networking, and other enabled OCI services might generate separate usage. Estimate usage for those resources separately from deployment Compute.

A monthly estimate isn't a commitment. Actual usage depends on how many replicas run and how long they run.

Steps

  1. On the Applications list page, select the application that you want to create a deployment for. If you need help finding the list page, see Listing applications.
  2. On the application's detail page, select the Deployments tab.
  3. Select Create deployment.
  4. Select the following container image information.
    • repository compartment
    • repository
    • version tag
  5. (Optional) To make this deployment the running version for the application, enable Deploy and activate.

    Only one deployment can be active at a time. When you activate a deployment, the previous active deployment becomes inactive.

  6. Select Create deployment.
    Note

    Before you use the deployment, wait until the status is Active and the container is up and ready to serve requests.

Tip

After you create a deployment, you can view its details and perform other tasks, such as activating it, copying its OCID, or deleting it. Use the Actions menu (three dots) in the Console to access these tasks. For more information, see Listing deployments.