Respect retryDelay in Gemini 429 Responses for Smoother User Experience

[fadenb]

Hi RooCode team,

I really enjoy using RooCode - awesome work!

When working with the Google Gemini API, I noticed that RooCode uses a generic exponential backoff when it hits a 429 error. However, the API response includes a retryDelay field that tells the client exactly how long to wait.

It would be great if RooCode could parse and use that retryDelay instead. I think it would improve the user experience while following the guidance the provider gives to the client.

Here’s an example response:

[429 Too Many Requests] You exceeded your current quota, please check your plan and billing details. 
For more information on this error, head to: https://ai.google.dev/gemini-api/docs/rate-limits. 

[{"@type":"type.googleapis.com/google.rpc.QuotaFailure","violations":[
  {"quotaMetric":"generativelanguage.googleapis.com/generate_requests_per_model",
   "quotaId":"GenerateRequestsPerMinutePerProjectPerModel",
   "quotaDimensions":{"location":"global","model":"gemini-2.0-pro-exp"},
   "quotaValue":"5"}
]},
{"@type":"type.googleapis.com/google.rpc.Help",
 "links":[{"description":"Learn more about Gemini API quotas",
           "url":"https://ai.google.dev/gemini-api/docs/rate-limits"}]
},
{"@type":"type.googleapis.com/google.rpc.RetryInfo", "retryDelay":"5s"}
]

Thanks for considering this!

P.s. "gemini-2.0-pro-exp" is part of the API error response even though 2.5 is used 🤷

[mrubens]

Can you try in 3.11.0 and reopen if still an issue? Thanks!

[fadenb]

Tested with 3.11.0 and it works exactly as I would have hoped ✅
Thank you very much!