Fallback Retry Policy

The Fallback Retry Policy API configures how the Allternit LLM gateway retries failed requests and whether it sends a cross-provider fallbackModels chain to the Gizzi runtime. The actual retry orchestration lives in the gateway failover layer; this API owns the durable configuration. All fallback retry policy endpoints require the caller to be an organization owner or admin.

Base URL

https://api.allternit.com/api/v1/admin/fallback-retry-policy

For local development:

http://127.0.0.1:8013/api/v1/admin/fallback-retry-policy

Endpoints

Method	Path	Description
`GET`	`/api/v1/admin/fallback-retry-policy`	Read the policy
`PUT`	`/api/v1/admin/fallback-retry-policy`	Update the policy

Policy fields

Field	Type	Description
`enabled`	boolean	Whether retries and fallback chains are enabled
`max_retries`	integer	Maximum retries per request (0–5)
`retryable_statuses`	string[]	Statuses that trigger a retry
`retryable_errors`	string[]	Error types or `*` for all errors
`base_delay_ms`	integer	Initial retry delay in milliseconds
`max_delay_ms`	integer	Maximum retry delay in milliseconds
`fallback_chain_enabled`	boolean	Whether to send `fallbackModels` to the runtime

Retryable statuses: refusal, error, rate_limited, budget_exceeded, dlp_blocked, client_disconnected, timeout.

Get the policy

curl https://api.allternit.com/api/v1/admin/fallback-retry-policy \
  -H "Authorization: Bearer $ADMIN_TOKEN"

Response

{
  "org_id": "org_01hqx8x7x7x7x7x7x7x7x7x",
  "enabled": true,
  "max_retries": 2,
  "retryable_statuses": ["refusal", "error", "rate_limited", "timeout"],
  "retryable_errors": ["*"],
  "base_delay_ms": 500,
  "max_delay_ms": 8000,
  "fallback_chain_enabled": true,
  "created_at": "2026-08-13T18:00:00Z",
  "updated_at": "2026-08-13T18:00:00Z"
}

Update the policy

curl -X PUT https://api.allternit.com/api/v1/admin/fallback-retry-policy \
  -H "Authorization: Bearer $ADMIN_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "enabled": true,
    "max_retries": 3,
    "retryable_statuses": ["refusal", "error", "rate_limited", "timeout"],
    "retryable_errors": ["rate_limit", "server_error"],
    "base_delay_ms": 1000,
    "max_delay_ms": 16000,
    "fallback_chain_enabled": true
  }'

Response

{
  "updated": true
}

Error codes

Status	Code	Description
`400`	`invalid_max_retries`	`max_retries` must be between 0 and 5.
`400`	`invalid_statuses`	`retryable_statuses` must be a non-empty array.
`400`	`invalid_status`	One or more statuses are not retryable.
`400`	`invalid_delay`	`base_delay_ms` or `max_delay_ms` must be positive.
`403`	`organization_required`	Caller is not associated with an organization.
`403`	`insufficient_role`	Caller is not an organization owner or admin.
`500`	`internal_error`	Database or server error.

Fallback Credits — credit primary-provider failures resolved by fallbacks
Admin Analytics — monitor request volume and status breakdowns
Latency Budgets — set and monitor model latency budgets

Get Started

Core Concepts

Gizzi Code CLI

Tools & ACI

Security & Enterprise

Admin API

BYOC Runtime

Surfaces

CI CD

Guides

Cookbooks

Release Notes

Fallback Retry Policy

Fallback Retry Policy

Base URL

Endpoints

Policy fields

Get the policy

Response

Update the policy

Response

Error codes

​Fallback Retry Policy

​Base URL

​Endpoints

​Policy fields

​Get the policy

​Response

​Update the policy

​Response

​Error codes

​Related pages

Fallback Retry Policy

Base URL

Endpoints

Policy fields

Get the policy

Response

Update the policy

Response

Error codes

Related pages