specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: Elastic Observability
providerId: elastic-observability
created: '2026-05-04'
modified: '2026-08-29'
generated: '2026-08-29'
method: searched
source: https://www.elastic.co/docs/reference/opentelemetry/motlp/rate-limiting
docs:
- https://www.elastic.co/docs/reference/opentelemetry/motlp/rate-limiting
- https://www.elastic.co/docs/reference/cloud/cloud-hosted/ec-api-rate-limiting
replaces: >-
  A bulk-sweep scaffold dated 2026-05-04 that asserted 10/100/1000 requests-per-minute tiers and
  100,000 monthly request quotas. None of those numbers are published by Elastic, and the product
  is not metered per request. Replaced wholesale on 2026-08-29.
limit_count: 0
limit_count_note: >-
  Elastic publishes NO numeric rate limit for the Observability intake API. This is an honest zero,
  not an unchecked one: the docs state that ingest capacity is a function of the deployment
  ("Elastic Cloud Hosted: limited by your Elasticsearch cluster capacity"; "Elastic Cloud
  Serverless: Elastic manages scaling automatically") rather than a published per-key ceiling. The
  runtime signal below is what an agent actually gets.
runtime_signals:
- surface: Elastic Managed OTLP endpoint (/v1/traces, /v1/metrics, /v1/logs and the OTLP gRPC
    Export paths)
  status_on_exhaustion: 429
  grpc_status: 'rpc error: code = ResourceExhausted desc = request exceeded available capacity'
  headers: []
  headers_note: >-
    No Retry-After and no X-RateLimit-*/RateLimit-* header is documented for this surface.
  scope: per deployment / per project capacity
  source: https://www.elastic.co/docs/reference/opentelemetry/motlp/rate-limiting
- surface: Elastic Cloud control-plane API (api.elastic-cloud.com) — adjacent surface, not the
    intake API
  status_on_exhaustion: 429
  headers:
  - x-ratelimit-limit
  - x-ratelimit-interval
  - x-ratelimit-remaining
  headers_note: >-
    Elastic documents these three response headers verbatim: "API calls are rate limited in a
    timing window. The current remaining available calls quota is available through the following
    header fields, included in each API call response." The window duration and the numeric ceiling
    are not published.
  scope: per user, per second (the window value itself is returned in x-ratelimit-interval)
  source: https://www.elastic.co/docs/reference/cloud/cloud-hosted/ec-api-rate-limiting
limits: []
policies:
- name: Backoff
  description: >-
    On 429 / ResourceExhausted, OpenTelemetry exporters and Elastic APM agents retry with
    exponential backoff. Because the intake API is append-only with no idempotency key, a retry
    after an ambiguous failure can duplicate telemetry — see
    conventions/elastic-observability-conventions.yml.
- name: Capacity, not quota
  description: >-
    On Elastic Cloud Hosted, ingest throughput is bounded by provisioned cluster capacity; on
    Serverless, Elastic scales automatically and bills per GB. Neither model exposes a
    requests-per-minute quota a client can read ahead of time.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com