Rate limits now count per IP
The published global ceiling drops to 1000 req/60 s to match what was already enforced. Strict and Global count per IP, not per credential.
ChangelogBreaking
The documented global ceiling goes from 2000 to 1000 requests per 60 seconds. This is not a reduction — the API was already enforcing 1000 — so nothing changes for a running integration.
What else changed
- Global: 1000 req / 60 s — the published figure now matches what an edge backstop was already enforcing
- Strict and Global count per originating IP, not per credential: several keys on one host share those budgets
- Standard still counts per credential — per API key, or per client for OAuth2 and dashboard sessions
/v1/accounts/importsjoins the Standard tier (300 req / 60 s)
Where to go next
Three filters returned the wrong rows
A literal `%` in a filter value dropped the filter. `series_code` matched any series containing the code. Sorting customers ignored the documented field names. **Only `series_code` changes a response you may rely on today.**
Request bodies limited to 2 MB
Bodies above 2 MB answer `413 REQUEST_BODY_TOO_LARGE`. `411 LENGTH_REQUIRED` is gone — chunked encoding is accepted.