ποΈ Overview
EGS exposes a number of APIs to create and manage slices and GPU provision request (GPRs)
ποΈ Authentication
The client needs to authenticate using the API Token to get a bearer token for the API calls.
ποΈ API Key
This topic describes the REST APIs to manage API keys.
ποΈ Clusters
This topic describes the API used to get the list of clusters registered with the EGS Controller.
ποΈ Workspace
A workspace is associated with one or more namespaces and a user.
ποΈ Workspace Policies
This topic describes APIs to list, retrieve, and update workspace policies.
ποΈ Inventory
List GPU nodes available for allocation requests (GPRs) in the cluster. These nodes are managed and allocated by EGS.
ποΈ GPR Templates
This topic describes the REST APIs to manage GPR Templates.
ποΈ GPR Template Binding
A GPR template binding is the GPR template that associates with the workspace.
ποΈ GPR
A GPR is a GPU Provisioning request to EGS to allocate GPUs to a workspace for a requested duration.
ποΈ GPU Availability
The GPU availability is determined by the current state of the GPU resources in EGS. EGS continuously monitors GPU usage and
ποΈ Inference Endpoints
An Inference Endpoint is a hosted service to perform inference tasks such as
ποΈ Workload Placement
This topic describes APIs used to create, get, update, and delete a Workload Placement.