How many parallel requests can it handle? (Highload ready)

Feature Description

Is there any analytics available on how many parallel requests Morphic can handle from users? What are the architectural limitations in terms of: а) back-end b) front-end

Are queues implemented? If possible, could you share your experience testing under high user load?

Use Case

Where is the bottleneck in terms of components and tools? Which will fail first—web search or the LLM service—under high load, say, 10 requests per second?

Additional context

No response