Role, permissions and rate limitings

Roles define what actions users can perform within the system through permissions, and what resource limits apply through rate limiting.

Available permissions

Permission	Description
admin	Full administrative access
create_public_collection	Create public collections
read_metric	Read prometheus /metrics endpoint
provide_models	Provide models to the system

Rate limiting

Rate limiting controls model usage for users with a specific role. Each limit has three components defined in the limits parameter:

• model: The model name
• type: The limit type
• value: The limit value (if null, the limit is not applied)

tip

Rate limiting allow to control model access. If a limit is set to 0, the model will be inaccessible.

Available limit types

Limit Type	Description
tpm	Tokens per minute
tpd	Tokens per day
rpm	Requests per minute
rpd	Requests per day

Example:

{
  "model": "my-language-model",
  "type": "tpm",
  "value": 100000
}

info

Rate limiting requires Redis to be configured. For more information about Redis setup and rate limiting strategies, see Redis documentation.

Managing roles

Create role

curl -X POST http://localhost:8000/v1/admin/roles \
  -H "Authorization: Bearer <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Developer",
    "permissions": ["create_public_collection"],
    "limits": [
      {
        "model": "my-language-model",
        "type": "tpm",
        "value": 100000
      }
    ]
  }'

Get roles

curl -X GET http://localhost:8000/v1/admin/roles \
  -H "Authorization: Bearer <api_key>"

Get role by ID

curl -X GET http://localhost:8000/v1/admin/roles/1 \
  -H "Authorization: Bearer <api_key>"

Update role

curl -X PATCH http://localhost:8000/v1/admin/roles/1 \
  -H "Authorization: Bearer <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Senior Developer",
    "permissions": ["create_public_collection", "read_metric"]
  }'

Delete role

curl -X DELETE http://localhost:8000/v1/admin/roles/1 \
  -H "Authorization: Bearer <api_key>"