Hey everyone, I’ve run into a really strange issue with the Platform API returning 429 Too Many Requests immediately when ramping JMeter threads to 200 concurrent users on /api/v2/conversations. The rate limit header says 60 req/min but docs suggest higher for this endpoint.
Using Genesys Cloud EU1 and standard bearer token auth. JMeter config is simple GET requests with 10s pause. Is there a hidden tenant-level cap for load testing scenarios?
To fix this easily, this is to adjust the request pacing in your jmeter script. the platform api enforces strict rate limits per tenant, and hitting /api/v2/conversations with 200 concurrent threads immediately triggers the 429 response. even with a 10s pause, the initial burst exceeds the 60 req/min threshold. consider implementing a constant throughput timer in jmeter to limit the requests to 55 per minute. this ensures steady state testing without triggering the circuit breaker. additionally, verify if the bearer token has the correct scopes for conversation read access. sometimes, insufficient permissions cause unexpected throttling behavior. if you are testing specific conversation details, ensure the query parameters are optimized to reduce payload size. this approach aligns with standard performance testing practices for genesys cloud apis. monitoring the rate limit headers in the response can provide further insights into the remaining quota.
TL;DR: Zendesk doesn’t use this model.
This is caused by the tenant-level circuit breaker kicking in before the 10s pause takes effect. Switch JMeter to a Constant Throughput Timer capped at 50 req/min to avoid the burst.