Alauda AI

Introduction

Connections let you store reusable access settings for external model sources and data services at the project level. In Alauda AI, a connection is stored as a Kubernetes Secret in the project namespace, and a connection type is stored as a Kubernetes ConfigMap in the kube-public namespace.

TOC

Core ConceptsConnectionConnection TypeSupported Connection Kinds

Core Concepts

Connection

A connection stores the actual access parameters used by workloads, such as:

  • S3 access key, secret key, endpoint, region, and bucket
  • OCI registry host and registry authentication
  • A direct remote URI

Connection Type

A connection type defines the form that users see when they create or edit a connection. It controls:

  • Which fields are displayed
  • Which fields are required
  • Default values and read-only settings
  • Whether the resulting connection is compatible with URI, OCI, or S3 model serving

Supported Connection Kinds

KindWhat the connection storesHow model deployment uses it
URIA full remote URI in the URI fieldThe deployment form reads the URI directly from the connection
OCI-compliant registryRegistry host, registry credentials, and access modeThe deployment form uses the connection secret as an image pull secret and asks you for the relative model path, such as repository:tag
S3-compatible object storageAccess key, secret key, endpoint, optional region, and bucketThe deployment form uses the bucket from the connection and asks you for the object path under that bucket

For step-by-step operations, see Using Connections.