Rate limits and pacing
WhaBuzz does not rate limit the API in v1; Meta's limits pass through. How to pace sends and read timing headers.
#WhaBuzz does not throttle your API calls
There is no WhaBuzz request quota on /v1 in v1. The only WhaBuzz rate limit is on sign-up, login and password-reset in the console (429 rate_limited with Retry-After).
#Meta's limits pass through
Meta limits you, and its errors reach you unchanged with "source":"meta". The ones you will meet:
| Meta code | Meaning | What to do |
|---|---|---|
130429 | Throughput limit reached (messages per second per number) | Back off and retry later. Pace sends. |
131056 | Too many messages to the same recipient in a short time | Wait before sending that customer again. |
131048, 131049 | Spam rate or marketing frequency limits | Do not retry the same message; review content and audience. |
A messaging tier (how many unique customers you may start conversations with per day) applies to each number and is reported in number.quality events. The full list is on Meta's error page.
#Pacing bulk sends
Use batches and let WhaBuzz pace for you: 20 messages per second per standard number, 50 for a high-throughput number. A scheduler runs every minute and sends at most 900 items per run, so plan on about 900 messages per minute per run. Failed items are not retried; list them with ?status=failed and resend what you choose.
If you send one at a time instead, keep to the same ceiling and treat 130429 as a signal to slow down.
#Reading latency
Every response carries Server-Timing: wb;dur=12, meta;dur=184 and X-WhaBuzz-Processing-Ms. wb is time spent in WhaBuzz; meta is time waiting on Meta's Graph API. WhaBuzz gives Meta 15 seconds, then answers 502 meta_unreachable (outcome unknown; see Idempotency).
This page as Markdown: /guides/rate-limits.md. All docs in one file: /llms-full.txt.