generated: '2026-08-29'
method: searched
source: https://docs.tailor.tech/reference/platform/platform-limits
docs: https://docs.tailor.tech/reference/platform/platform-limits
name: Tailor platform limits
description: >-
  Tailor publishes a single Platform Limits reference covering the one true rate limit
  (incoming webhooks) plus the concurrency, memory and payload ceilings that behave like
  limits from a caller's perspective. It also publishes a separate Timeouts reference.
  The rate limit is the only one that returns a 429; the concurrency caps queue instead.
limit_count: 11
response_headers:
- name: RateLimit-Limit
  description: The limit applied to the current window.
- name: RateLimit-Remaining
  description: Requests remaining in the current window.
- name: RateLimit-Reset
  description: Seconds until the current window resets.
header_style: IETF draft (unprefixed RateLimit-*), not X-RateLimit-*
exhaustion_status: 429
exhaustion_behavior: >-
  Rate-limited requests are REJECTED, not queued, at the API gateway before the executor
  is invoked. There is no platform-side retry.
rate_limits:
- name: Incoming webhook request rate
  scope: per-workspace
  window: 1 second (fixed, not sliding)
  limit: 100
  unit: requests
  burst: null
  status_on_exhaustion: 429
  headers: [RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset]
  note: >-
    Counted per workspace across all incoming webhook endpoints, not per executor and
    not per source IP. Because windows are fixed rather than sliding, sustained traffic
    at exactly 100 req/sec still sees some 429s at window boundaries — the docs
    explicitly say to target below the limit.
concurrency_limits:
- name: Workflow — workspace concurrent executions
  scope: per-workspace
  limit: 50
  behavior: excess executions stay PENDING and are re-evaluated each scheduler tick; never rejected
- name: Workflow — per-workflow concurrent executions
  scope: per-workflow
  limit: 20
  behavior: excess executions stay PENDING
- name: Workflow — workspace concurrent job function dispatches
  scope: per-workspace
  limit: 100
  behavior: dispatch suspended, workflow moves to PENDING_RESUME, retried after ~5s
- name: Workflow — per-key concurrent job functions (platform fallback)
  scope: per executionPolicyKey
  limit: 50
  behavior: only applies when the matching execution policy declares no concurrencyPolicy.maxConcurrentExecutions
- name: Executor — workspace concurrent job function operations
  scope: per-workspace
  limit: 100
  behavior: >-
    queued oldest-first, never dropped. Docs give the drain formula:
    (queued ÷ 100) × average execution duration.
- name: Recursive call depth
  scope: per-request-chain
  limit: 10
  behavior: request rejected with BadRequest beyond 10 levels of nested platform-to-platform calls
payload_limits:
- name: Function memory
  limit: 32 MB
  behavior: execution terminated with "Memory limit exceeded"
- name: JobFunction memory
  limit: 256 MB
  behavior: execution terminated with "Memory limit exceeded"
- name: TailorDB select result size
  limit: 128 MB
  behavior: query fails; paginate or filter columns
- name: Fetch buffered response body
  limit: 10 MB
  behavior: '.text()/.json() fail; streaming reader methods are exempt'
timeouts:
  source: https://docs.tailor.tech/reference/platform/timeouts
  values:
  - service: API gateway
    timeout: 60s
  - service: Resolver
    timeout: 60s
  - service: Executor (TargetTailorGraphql / TargetWebhook)
    timeout: 60s
    on_failure: retry up to 10 attempts
  - service: TailorDB hooks (PreHook/PostHook)
    timeout: 2s
  - service: Resolver hooks (PreHook/PostHook)
    timeout: 2s
  - service: Function
    timeout: 60s
  - service: JobFunction
    timeout: 24h
control_plane:
  documented_rate_limit: false
  note: >-
    No published rate limit for the OperatorService control-plane API. One RPC in the
    contract declares ResourceExhausted, but no numeric quota is documented.
x-evidence:
  fetched: '2026-08-29'
  probes:
  - url: https://docs.tailor.tech/reference/platform/platform-limits.md
    status: 200
  - url: https://docs.tailor.tech/reference/platform/timeouts.md
    status: 200