Looking for advice on handling rate limiting for analytics endpoints during load testing.
Running JMeter 5.6.2 from Singapore. Targeting Genesys Cloud platform_api. Script execution fails when threads hit 50. Getting 429 Too Many Requests on /api/v2/analytics/interactions/summary.
The goal is to simulate 50 concurrent agents pulling interaction summaries every 30 seconds. The JMeter test plan uses a Thread Group with 50 threads, ramp-up 10 seconds, loop count 100. Each thread sends a GET request to the analytics endpoint with a date range of the last hour.
At 20 threads, the test runs smoothly with 200 OK responses. At 50 threads, the first batch of requests succeeds, but subsequent requests return 429. The response headers include Retry-After: 1. I have added a 1-second pause between requests in JMeter, but the 429 errors persist.
Checking the Genesys Cloud documentation, the rate limit for analytics endpoints is 10 requests per second per org. With 50 threads sending requests every 30 seconds, the theoretical load is 1.67 requests per second, which is well under the limit. However, the 429 errors suggest that the API is seeing a higher rate or there is a different limit at play.
Is there a per-user rate limit that is being hit? Or is there a global limit that is being exceeded? How can I adjust the JMeter script to stay within the rate limits? I have tried using the Constant Throughput Timer in JMeter to limit the overall throughput to 10 requests per second, but the 429 errors still occur.
Any insights on how to properly load test analytics endpoints without hitting rate limits would be appreciated. I am trying to understand the true capacity of the API under concurrent load.