Rate Limits

Understand quotas, retry guidance, and headers returned by API v1.

Default Limits

ServoAgent enforces token-bucket rate limiting per API key, workspace, and endpoint class.

Burst capacity is available for short spikes, but sustained throughput follows your plan quota.

  • - Per API key baseline: 100 req/min
  • - Workspace burst window: 250 req/min
  • - Endpoint-specific limits for heavy operations

Response Headers

Every API response includes current rate state for adaptive clients.

X-RateLimit-Limit: 100
X-RateLimit-Remaining: 84
X-RateLimit-Reset: 1735689600
Retry-After: 9

Retry Strategy

On 429 responses, back off using Retry-After and apply jitter to avoid synchronized retries.

Write idempotent mutation requests where possible so retries are safe in network failure scenarios.