どのソースが決定を制御すべきですか？

Use 100 requests per minute – the most authoritative source in the packet is the official Harbor API reference (updated 20 Aug 2026).

What to verify before launch

1. Confirm the reference is still current – check that no newer version of the API documentation has changed the limit.
2. Test the rate‑limit enforcement – send bursts of requests from your integration to see how the API responds (e.g., 429 status, retry‑after header).
3. Check for endpoint‑specific or account‑type limits – some endpoints or user tiers may have different caps or burst allowances.
4. Determine the scope of the limit – is it per API key, per IP, or global?
5. Validate burst capacity – many APIs allow a short burst above the nominal rate; confirm the maximum burst size.
6. Ensure your integration’s traffic pattern stays within the limit – plan for peak usage and add back‑off logic if needed.

By following these checks, you’ll be confident that the 100 req/min limit is correct and that your integration will operate reliably.

The integration should use 100 requests per minute as the default limit, based on the official reference. Before launch, the team should verify if there have been any updates to the official documentation since August 20, 2026, and confirm the current limit with the vendor.