generated: '2026-09-13'
method: searched
source: https://es-pera.org/metodologia/api/ and openapi/es-pera-openapi.json (info.description + TooManyRequests response)
name: ES·pera API rate limits
scope: per-client (anonymous IP)
exhaustion_status: 429
limits:
  - name: global-anonymous
    scope: per-client
    window: 60s
    limit: 120
    applies_to: all operations
    note: Documented as "El tráfico anónimo está limitado a 120 solicitudes por minuto."
  - name: waiting-facets-observations
    scope: per-client
    window: 60s
    limit: 20
    applies_to:
      - getWaitingFacets   # GET /waiting/facets
      - listWaitingObservations  # GET /waiting/observations
    note: These two endpoints carry an additional stricter 20/min sub-limit within the global 120/min.
response_headers:
  - name: RateLimit-Limit
    description: Maximum applicable in the current window (example 120).
    observed_on: 429
  - name: RateLimit-Policy
    description: Policy applied with limit and window in seconds (example "120;w=60").
    observed_on: 429
  - name: Retry-After
    description: Seconds the client must wait before retrying.
    observed_on: 429
notes:
  - The documented RateLimit-* / Retry-After headers are declared on the 429 TooManyRequests
    response in the OpenAPI; a live 200 /api/v1/health response did not echo RateLimit-* headers.
  - Clients must honor Retry-After on 429 and avoid parallel retries (per the provider Agent Skill).