Quick question about… hitting 429 Too Many Requests on the POST /api/v2/conversations/messages endpoint while running a concurrent load test from Singapore. Setup involves JMeter 5.6.2 with 100 virtual users sending text messages to a single bot endpoint via the Platform API. The test plan uses a CSV Data Set Config for unique participant IDs and a constant throughput timer set to 10 samples/second per thread group. Initially, requests succeed with 201 Created responses, but after roughly 50 seconds, the error rate spikes to 40% with 429 status codes and Retry-After headers suggesting a 10-second wait. The x-gc-request-id headers are unique, and OAuth tokens are refreshed every 50 minutes to avoid expiration issues. Is this a hard limit for the messaging API, or is there a specific header or pagination strategy that helps distribute the load better? Also, wondering if the WebSocket connection limit for the analytics event streaming functionality is being hit simultaneously, causing cascading failures. Any insights on the actual rate limit thresholds for bulk message ingestion would be appreciated.
2 Likes