Categories

Versions

Manage the endpoint infrastructure

Before the arrival of a container-based architecture, the scoring agent was a free-standing entity; now, depending on your endpoint infrastructure, it may be standalone or it may be packaged within a Docker image.

When managing the endpoint infrastructure, recall that configuration occurs at multiple levels, similar to the situation for Altair AI Hub.

  1. The landing page

    On Altair AI Hub, there is an Endpoints landing page for Web API endpoints.

  2. The scoring agent's REST API

    The complete REST API documentation of the Scoring Agent is available as OpenApi 3.0 specification and is published on SwaggerHub.

  3. The platform's environment file

    If your scoring agent is containerized, you can configure it via the environment file for docker-compose.

    See the table of system settings for

  4. Within containers

    If your scoring agent is containerized, and you want to modify its files, you must first get access to the container.

The landing page

Web API endpoints

From the landing page of Altair AI Hub, select Projects and locate the Endpoints page:

../create/img/endpoint-deployed.png

Note that this page only displays the Web API endpoints provided by Altair AI Hub's Web API agents. Regular endpoints, provided by regular scoring agents, are not included in this list.

Within this list, you can take Endpoint Actions:

as well as Deployment Actions -- note however that when you edit or delete a deployment, your action affects every endpoint connected with that deployment.

If you only wish to edit or delete one specific endpoint, not the whole deployment, you must

  1. first edit the deployment, then
  2. edit or delete the endpoint in question.

../create/img/endpoint-setup-save-and-deploy.png

Editing files within containers: the home directory

When the scoring agent is containerized and you want to make changes, you must first get access to the container.

  1. Find the container running scoring-agent:

     $ docker container ls | grep rapidminer-scoringagent
    
     CONTAINER ID   IMAGE
     63b24f5fb5c9   rapidminer/rapidminer-scoringagent:10.1.3
    
  2. Enter the running container:

     $ docker exec -it <CONTAINER_ID> bash
    
  3. Within the container, find the home directory

     rapidminer@scoring-agent:/$ ls scoring-agent/home
    
     config  deployments  log  pid  resources  tmp
    
  4. Edit the file of interest with an available editor, such as nano or vim.

  5. Exit the container:

     rapidminer@scoring-agent:/$ exit
    
  6. Restart scoring-agent:

     $ docker-compose restart scoring-agent