Rate Limits

Nova throttles requests per credential to protect the platform.

Default quotas

Limits apply per credential, not per IP. Spread bursts across multiple
keys if you need more headroom (or ask us to raise the quota).

Response headers

Every response includes:

X-RateLimit-Limit: 100
X-RateLimit-Remaining: 87
X-RateLimit-Reset: 1714214140

X-RateLimit-Reset is a unix timestamp — the next time the bucket
refills.

When you hit the limit

Status 429 with body:

{
  "error": {
    "code": "RATE_LIMITED",
    "message": "Too many requests",
    "details": { "retryAfterSeconds": 7 }
  }
}

The Retry-After header is also set. Honour it. Continued hammering can
escalate to a temporary block.

Best practice

• Cache reads where you can — partner trees, product catalogues, and
  config rarely change.
• Use webhooks instead of polling.
• Use bulk endpoints where they exist (e.g. SIM allocation) instead of
  per-record loops.
• For backfills or migrations, talk to us first — we can usually raise
  the limit temporarily.