Request for Guidance on NVIDIA NIM Rate Limits for Multi-Agent Testing (original) (raw)

Hello NVIDIA Team,

I am currently evaluating NVIDIA NIM models through build.nvidia.com while prototyping a multi-agent workflow.

My account currently appears to be limited to approximately 40 requests per minute. During testing, some agent workflows generate short bursts of requests that occasionally trigger HTTP 429 rate limit responses, despite implementing throttling and retry logic.

I would appreciate clarification on the following:

1. Are higher rate limits available for developers actively testing applications on build.nvidia.com?
2. If so, what is the recommended process for requesting an increase?
3. Are there recommended deployment options for users whose workloads exceed the standard development limits?

Current usage is strictly for development and testing purposes.

Thank you for any guidance you can provide.

Best regards,

Carlos Villanueva