generated: '2026-09-09'
method: probed
source: https://hq.advancednavigation.com.au/kinematica/api/alldatasets?v=1.2
docs: https://www.advancednavigation.com/downloads/versioned/kinematica/kinematica_api_reference_manual-1.6.pdf
limit_count: 1
note: >-
  Published in section 7 of the Kinematica API Reference Manual (added in document v1.6 / server
  v1.5.22, 06/01/2026) AND observed live on an unauthenticated response on 2026-09-09 — the response
  headers below were read off the wire, not only off the page.
rate_limits:
- name: Kinematica API global limit
  scope: per-api-key AND per-client-ip (dual bucket)
  limit: 6000
  unit: requests
  window: 1 minute
  burst: null
  refill: tokens refill every minute
  applies_to: all /api endpoints
  enforcement: >-
    "Dual bucket: one per IP. One per API key. Request is denied if either bucket is empty."
  exhaustion_status: 429
  docs: https://www.advancednavigation.com/downloads/versioned/kinematica/kinematica_api_reference_manual-1.6.pdf
headers:
- name: X-RateLimit-Limit
  meaning: total token limit for the policy (observed value 6000)
  observed: true
- name: X-RateLimit-Remaining
  meaning: tokens remaining in the most restrictive bucket
  observed: true
- name: X-RateLimit-Remaining-IP
  meaning: tokens remaining in the per-IP bucket
  observed: true
- name: X-RateLimit-Remaining-User
  meaning: tokens remaining in the per-API-key bucket
  observed: true
- name: X-RateLimit-Reset
  meaning: unix timestamp when the window resets
  observed: true
- name: Retry-After
  meaning: seconds to wait before retrying; documented on the 429 response only
  observed: false
- name: Cache-Control
  meaning: "no-store on a 429, so a rate-limited response is never cached"
  observed: false
exhaustion_body: |
  {
      "success": false,
      "message": "Too many requests",
      "retryAfter": 60
  }
x-evidence:
  fetched: '2026-09-09'
  url: https://hq.advancednavigation.com.au/kinematica/api/alldatasets?v=1.2&apiKey=invalid
  http_status: 200
  response_headers:
    X-Ratelimit-Limit: '6000'
    X-Ratelimit-Remaining: '5998'
    X-Ratelimit-Remaining-Ip: '5998'
    X-Ratelimit-Remaining-User: '5999'
    X-Ratelimit-Reset: '1788990535'
  note: >-
    Rate-limit headers are returned on ordinary 2xx responses, not only on 429 — an agent can read its
    remaining budget from every call. Both bucket counters are exposed separately, which is unusual
    and useful: a client can tell whether it is the IP or the key that is running out.