⢠LiveKit suggests starting with 4 cores and 8GB of memory for every 25 concurrent sessions in most voice-to-voice applications.
⢠In a real-world test, 30 concurrent participants on a 4-core, 8GB machine used approximately 3.8 cores and 2.8GB of memory at peak.
Hey there! I totally get you're trying to figure out the right resources for your outbound calls. From what I can see, handling 10 parallel calls should be pretty manageable. These LiveKit recommendations are super helpful ā looks like you'll probably need less than 4 cores and 8GB of memory. But hey, every setup is unique! If you want to dial in the perfect configuration for your specific needs, I'd love to hear more details and help you nail down the exact resources. Just shoot over some more info, and we'll figure this out together. š Sources:
Deploying to production | LiveKit Docs