Skip to content
Error codes · Rate limits, quotas & billing

RU-2003: API write rate limit exceeded

A key spends one unit of a 30-a-minute write bucket on every request that changes data, and a Free account's non-GET requests also spend from the shared 60-a-minute account budget. The first of the two to run out is the one that refuses.

Where it is raised

  • Checks: 6 routes
  • Status Pages: 6 routes
  • Notification Channels: 4 routes
  • Incidents: 2 routes
  • Maintenance: 2 routes
  • Errors: 3 routes
  • Agents: 3 routes
  • Keyed MCP endpoint, write tools

What triggers it

Every key has a write bucket of 30 requests a minute, and a Free account's shared 60 a minute applies to its non-GET requests as well. It is a separate bucket from the read one, so a key can be out of write budget and still read freely.

Example response

POST /api/v1/incidents
HTTP 429
Retry-After: <seconds>

{
  "error": "Rate limit exceeded. Try again shortly.",
  "code": "RU-2003"
}

Common causes

  • Bulk-creating or bulk-updating resources in a fast loop.
  • A sync job replaying its whole state on every run instead of diffing.
  • Several write integrations on a Free account drawing on the one 60-a-minute account budget.

How to fix it

  1. Honor the Retry-After header and spread bulk operations out.
  2. Only write what changed: diff before you sync.

Related codes