v0.2.4 - Rate Limiting Fix
What's Fixed
Proper Rate Limiting for GitHub Models API
Previously, the plugin would blast embedding requests at the API, causing 429 (Too Many Requests) errors on large codebases.
Changes:
- Enforces 15 requests/minute limit (1 request per 4 seconds)
- Sequential processing (concurrency reduced from 3 to 1)
- Exponential backoff on rate limit errors (10s → 20s → 40s → 60s max)
- Longer retry delays (minimum 5s, max 60s)
Impact:
- Indexing is slower but reliable
- Large codebases (10k+ files) will take longer but won't fail with 429 errors
- For faster indexing on large codebases, consider using Ollama locally
Full Changelog
- fix: add proper rate limiting for GitHub Models API (15 req/min)