Need some troubleshooting help with a specific failure pattern when initiating screen recordings via api during high concurrency.
context:
region: us1
jmeter: 5.6
genesys cloud sdk: 2.3.0
flow: architect flow triggering /api/v2/interactions/screen-recordings
the issue occurs when ramping up to 50 concurrent agents. the api returns http 500 internal server error with message “unable to start recording service”. this happens roughly 12% of the time.
jmeter config:
- thread group: 50 threads
- loop count: 5
- ramp-up: 10s
- http request sampler pointing to /api/v2/interactions/screen-recordings
error response:
{
"message": "unable to start recording service",
"errors": [
{
"code": "500",
"message": "internal server error"
}
]
}
is there a known limit on screen recording initiation rate? or is this a capacity issue in us1? looking for best practices to handle this in load tests.