v0.2.2 - Rate Limit Handling

Bug Fixes

Improved Rate Limit Handling

When the embedding API returns 429 (Too Many Requests), the plugin now:

• Detects rate limits and shows a clear message instead of [object Object]
• Reduces concurrency from 3 parallel requests to 1
• Uses exponential backoff (factor 2, max 30 seconds) between retries
• Retries gracefully instead of failing immediately

Better Error Messages

• Fixed error display showing [object Object] instead of actual error text
• Rate limit errors now show: Rate limited (attempt X/Y): waiting before retry...

Notes

If you're hitting rate limits with GitHub Copilot's embedding API, the plugin will now automatically slow down and retry. For large codebases, consider:

1. Running /index during off-peak hours
2. Using the estimateOnly flag first to check the size
3. Configuring fewer retries in .opencode/codebase-index.json if needed