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: Mezmo
providerId: mezmo
created: '2026-05-25'
modified: '2026-05-25'
reconciled: false
description: |
  Best-effort capture of Mezmo's rate-limit posture. Mezmo enforces ingestion-rate and
  request-rate limits but does not publish a single canonical limits table. Verify with
  support for contract-specific numbers.
sources:
  - https://docs.mezmo.com/log-analysis-api
  - https://docs.mezmo.com/pipeline-api
limits:
  - api: mezmo-log-ingestion-api
    scope: per-account
    policy: token-bucket
    description: |
      Log ingestion accepts batched JSON. Practical limits: keep individual batch payloads
      under 10 MB. Sustained ingestion is bounded by the contracted ingest volume; bursts
      above the contracted rate may be queued, throttled with 429, or dropped depending on
      plan.
    entries:
      - label: Batch payload size
        limit: 10485760
        unit: bytes
        timeFrame: request
      - label: Lines per batch (recommended)
        limit: 10000
        unit: lines
        timeFrame: request
      - label: Sustained ingest rate
        limit: 'contract'
        unit: bytes/sec
        timeFrame: second
        description: Bounded by contracted volume; check usage/v2 for actual consumption.
  - api: mezmo-log-export-api
    scope: per-account
    policy: pagination
    description: |
      V1 export is plan-bounded for total result size. V2 export is paginated and rate-limited
      to protect storage; clients should respect 429s and continuation cursors.
    entries:
      - label: V2 export page size
        limit: 10000
        unit: lines
        timeFrame: request
  - api: mezmo-pipeline-api
    scope: per-account
    policy: token-bucket
    description: |
      Control-plane rate limit. CRUD against pipelines, sources, processors, destinations.
      Default control-plane RPS is generous; bulk operations should batch where possible.
    entries:
      - label: Control plane requests per minute
        limit: 600
        unit: requests
        timeFrame: minute
        description: Indicative — verify with support.
  - api: mezmo-pipeline-classification-api
    scope: per-account
    policy: token-bucket
    description: |
      AI-assisted classification calls are more expensive; expect lower per-minute caps and
      possible 429s under burst load.
    entries:
      - label: Classification requests per minute
        limit: 60
        unit: requests
        timeFrame: minute
        description: Indicative — verify with support.
notes:
  - Mezmo returns standard HTTP 429 when limits are exceeded; clients should retry with
    exponential backoff and respect Retry-After when present.
  - Suspension state (ingestion suspended) returns 403 on the ingestion endpoint until
    /v1/config/ingestion/resume is called.