Rate Limits

Laracord uses kyzegs/guzzle-rate-limit-middleware with Discord's dynamic bucket hashes and major parameters.

• Per-route state persists through Laravel cache.
• Laravel atomic locks serialize competing workers.
• Global capacity defaults to 50 requests per second per credential.
• Unexpected 429 responses honor headers or JSON retry_after.
• 401, 403, and non-shared 429 responses feed a configurable invalid-request safety budget.

Use Redis or another shared atomic-lock cache in multi-worker production deployments. All limits and safety margins are configurable under laracord.rate_limit.