Read upstream credential request limit
Generated from the checked-in source. Do not edit this reference by hand.
GET /api/v1/vendors/{vendor}/request-rate-limit
Platform administration only. Applies to mapped inference dispatches pinned to this saved credential across workspaces and gateway instances, independent of customer key RPM. Null is unlimited; zero denies dispatch. Revision starts at string 0 and survives credential edits. Rolling 60-second admissions count dispatch, including upstream failures and uncertain outcomes; polling and connection checks are not inference dispatch. Unlimited dispatches also count when a cap is subsequently enabled. No automatic route substitution or replay. Duplicated credentials saved under different vendor IDs have independent windows. Storage operations use bounded transactions: read-only snapshots for reads and a recording transaction for updates, with transaction-local 250 ms lock and 2 s statement limits. Exceeded waits return sanitized 503; failed updates do not advance policy/history. These bounds exclude authorization and pool acquisition.
Operation: getVendorRequestRateLimit.
Authentication
Section titled “Authentication”Use the credential and permission scope described above. Each object below is an alternative; schemes within one object are required together.
[ { "bearerAuth": [] }]HTTP request
Section titled “HTTP request”Set NIU_ORIGIN to your installation origin (without /v1). Replace path, query and header placeholders with the required values. Query examples use curl 7.87 or newer to encode values with --url-query. Use the credential scope described above; a bearer token must have the authority required by this operation.
curl --fail-with-body -X GET "$NIU_ORIGIN/api/v1/vendors/{vendor}/request-rate-limit" \ -H "Authorization: Bearer $NIU_BEARER_TOKEN"Parameters
Section titled “Parameters”vendor
Section titled “vendor”Location: path. Required.
{ "type": "string", "format": "uuid"}Request body
Section titled “Request body”No request body is declared.
Responses
Section titled “Responses”HTTP 200
Section titled “HTTP 200”Current policy or descending immutable history.
Content type: application/json.
{ "type": "object", "required": [ "data" ], "properties": { "data": { "type": "object", "additionalProperties": false, "required": [ "requests_per_minute", "revision" ], "properties": { "requests_per_minute": { "type": [ "integer", "null" ], "minimum": 0, "maximum": 1000000 }, "revision": { "type": "string", "pattern": "^[0-9]+$" } } } }}HTTP 400
Section titled “HTTP 400”Invalid limit, revision or query.
HTTP 401
Section titled “HTTP 401”Authentication required.
HTTP 403
Section titled “HTTP 403”Platform administration required.
HTTP 404
Section titled “HTTP 404”Credential configuration not found.
HTTP 409
Section titled “HTTP 409”Stale revision on policy update.
HTTP 503
Section titled “HTTP 503”Storage unavailable or bounded storage wait exceeded; no empty-success policy/history fallback.
Behavior and supported scope
Section titled “Behavior and supported scope”Platform administration only. Applies to mapped inference dispatches pinned to this saved credential across workspaces and gateway instances, independent of customer key RPM. Null is unlimited; zero denies dispatch. Revision starts at string 0 and survives credential edits. Rolling 60-second admissions count dispatch, including upstream failures and uncertain outcomes; polling and connection checks are not inference dispatch. Unlimited dispatches also count when a cap is subsequently enabled. No automatic route substitution or replay. Duplicated credentials saved under different vendor IDs have independent windows. Storage operations use bounded transactions: read-only snapshots for reads and a recording transaction for updates, with transaction-local 250 ms lock and 2 s statement limits. Exceeded waits return sanitized 503; failed updates do not advance policy/history. These bounds exclude authorization and pool acquisition.
Supported scope: implemented