please add per-mode Rate Limiting!

[C-Digi]

Personally I use Gemini API a lot, and their free tier (and 2.0 Pro on the paid tier) are rate limited at different rates. And other providers, like Openrouter which I use a lot as well (taking advantage of top-tier free models), don't have the same limits.
Please implement custom rate-limiting on a per-profile basis! This will especially be helpful for things like Orchestrator mode (mrubens shared discord mode) to switch between simple coding models, reasoning models, etc. which are attached to the different modes.
RooCode Rocks!
Thanks!

[mrubens]

This makes sense to me

[KJ7LNW]

Having a fallback "provider" could be nice, with an un-fallback timeout.

eg, if "copilot-sonnet-3.5" gives me a "rate limit" error, then I would love to specify "anthropic-sonnet-3.5" as a fallback, but only for 10 minutes and then it should retry "copilot-sonnet-3.5" again since copilot is fixed-cost instead of variable cost.

[KJ7LNW]

UI could have an option:

Fallback Provider [ Provider List Pulldown (v) ]	Fallback Timeout: [ time 600s]

[assaf-b]

Per mode, maybe even per model limiting, is critical. No matter how much you'd pay, some models are simply too rate limited to code, but still would be great for technical conversation and maybe architecture. Per mode or per model limiting is a must have feature in my eyes.

[C-Digi]

I agree, having it attached to each API profile would be best

[mrubens]

Soon!