Platform

API Gateway

Per-project rate-limit policies with a live bucket inspector.

▸ How it works

Every API key has a token bucket (refill rate + burst capacity). The bucket is held in ETS for O(1) atomic decrement and synced to Postgres every second. Policies define per-(project, key, route) limits. The dashboard shows the live state of every bucket — tokens remaining, refill rate — useful when debugging "why am I being rate-limited?".

▸ What this lets you do

✓ Protect upstream services from runaway clients
✓ Per-project, per-key, per-route policies — granular control
✓ Live bucket inspector — no guessing at rate-limit state
✓ Returns standard `429` + `Retry-After` headers

▸ Get it running

1 Default policies are sensible — no setup required
2 Override per-project in `API Gateway` → `Policies`
3 Inspect live buckets in the same page

▸ Code examples

Open `API Gateway`. Each ingest pipeline (ingest, RUM,
workers, AI, api-read) has its own card showing the live
bucket fill bar and current capacity / refill rate.

Click `Edit` on the card you want to override, then set:

  Burst capacity        5000
  Sustained rate / sec  500

Save — the override persists in `gateway_policies` and the
live bucket is reset to full so the new limit takes effect
on the next request. Click `reset` later to revert to the
pipeline default.

Where to find it

/app/p/:org/:project/gateway

Open in app →