generated: '2026-08-29'
method: derived
source: openapi/superb-ai-mlops-platform-openapi.json (429 responses on all 164 operations)
note: >-
  Superb AI declares that rate limiting EXISTS and is machine-detectable, but publishes no
  numbers and no runtime headers. Every one of the 164 operations in the MLOps Platform API
  declares a 429 response described as "Rate-limited (per-tenant token bucket)", and RATE_LIMITED
  and RESOURCE_LIMIT_EXCEEDED are values in the stable ErrorCode enum. No limit, window or burst
  figure is stated anywhere in the spec or on docs.superb-ai.com, and no X-RateLimit-*,
  RateLimit-* or Retry-After response header is declared. An agent therefore cannot pace itself:
  it discovers the ceiling only by hitting it, and has no published reset time to back off to.
limit_count: 0
algorithm: token bucket
scope: per-tenant (workspace)
source_of_truth: contract
headers:
  published: false
  x_ratelimit: false
  ratelimit_draft: false
  retry_after: false
  probe: >-
    Searched the full 536KB OpenAPI document for "Retry-After", "X-RateLimit", "RateLimit" —
    zero occurrences.
exhaustion:
  status_code: 429
  error_code: RATE_LIMITED
  envelope: '{"error": {"code": "RATE_LIMITED", "message": ..., "hint": ..., "details": ..., "request_id": ...}}'
  operations_declaring_429: 164
related_limits:
- name: JOB_PER_TENANT_CAP
  kind: concurrency cap
  detail: Enumerated error code — a per-tenant cap on concurrent jobs exists, but its value is unpublished.
- name: DATASET_ASSET_CAP_EXCEEDED
  kind: resource cap
  detail: Enumerated error code — a per-dataset asset cap exists, but its value is unpublished.
- name: RESOURCE_LIMIT_EXCEEDED
  kind: quota
  detail: Enumerated error code covering general quota exhaustion.
- name: PAYLOAD_TOO_LARGE
  kind: request size
  detail: Enumerated error code; no byte figure published.
limits: []