How should I properly to handle bulk intent updates via the Bot Platform API without triggering aggressive rate limiting?
Our AppFoundry integration processes dynamic intent updates from a client’s external knowledge base. The payload size varies, but we frequently process batches of 50-100 intents per minute.
The environment is a multi-org setup using OAuth 2.0. We are using the standard REST endpoints for intent management.
When the batch size exceeds approximately 30 intents within a 60-second window, the API returns a 429 Too Many Requests error. The response headers do not provide a clear Retry-After value. This breaks the synchronization logic in our premium app.
The documentation mentions rate limits for the platform API generally, but specific quotas for the Bot Platform endpoints remain vague. We need to ensure data consistency across multiple organizations.
Is there a recommended batching strategy or exponential backoff implementation specifically for intent training updates? We want to avoid hardcoding sleep intervals that might waste latency.
Any insights on the specific throttle limits for POST /api/v2/bots/{botId}/intents would be appreciated.