Feature request: Show rate limit usage and reset time in Copilot Chat UI

[fffokazaki]

Summary

When using Copilot Chat (especially with parallel/rapid requests), users can hit rate limits and only see a generic rate_limited error message.

Current message example:

• "Sorry, you've exceeded your rate limits... error code: rate_limited"

Problem

The current UI does not clearly show:

1. How close the user is to the limit (current usage vs quota)
2. When the limit will reset (exact timestamp or countdown)
3. Whether the limit is model-specific or global

Because of this, users cannot plan usage and repeatedly retry without knowing when it will recover.

Requested UX improvements

Please add a visible rate-limit status in Copilot Chat UI:

• Remaining quota / used quota (or a percentage)
• Estimated reset time (local timezone)
• Countdown until next available request window
• Scope clarification (per model, per time window, per plan)

Optional but helpful:

• A small status indicator near model selector
• Tooltip/Panel with detailed limits and docs link
• Retry recommendation based on actual reset window

Reproduction context

• Product: GitHub Copilot Chat in VS Code
• Scenario: Parallel or high-frequency requests
• Result: rate_limited shown, but no actionable visibility for timing/usage

Expected outcome

Users can predict and manage limits without trial-and-error, reducing frustration and unnecessary retries.

Thanks for considering this improvement.

[fffokazaki]

Additional context from user report (JP locale):\n\n- The error appears while asking Copilot Chat to do parallel tasks.\n- Message shown repeatedly: "Sorry, you've exceeded your rate limits... error code: rate_limited".\n- User cannot see current limit usage or when it will be lifted, so they keep retrying without actionable timing.\n\nUX requests (from real usage):\n1) Show current usage vs limit (or remaining requests)\n2) Show exact reset time in local timezone\n3) Show countdown until retry is likely to succeed\n4) Clarify whether the limit is per-model or account-wide\n\nThis would significantly reduce frustration and unnecessary retries.