@cloudflare/workers-utils@0.29.0
📦 cloudflare-workersView on GitHub →
✨ 5 features🔧 2 symbols
Summary
Wrangler now automatically retries '429 Too Many Requests' API errors and respects the 'Retry-After' header for more intelligent rate limit handling. The output file now includes 'retry_after_ms' for programmatic retry scheduling.
✨ New Features
- Wrangler commands now automatically retry '429 Too Many Requests' responses from the Cloudflare API, in addition to existing '5xx' error retries.
- When a retried Cloudflare API response includes a 'Retry-After' header, Wrangler will now wait for that specified duration before retrying, logging the wait time. Waits longer than 60 seconds will fail fast, providing the 'Retry-After' value for external scheduling.
- If a retryable API error is ultimately surfaced to the user, the error message will include a note with the 'Retry-After' duration.
- The 'command-failed' entry in the Wrangler output file now includes a 'retry_after_ms' field, allowing scripts and CI/CD pipelines to programmatically determine wait durations for retries.
- 'retryAfterMs' is now populated on APIErrors originating from direct R2 object requests, the Browser Rendering API, and errors surfaced from commands utilizing the official 'cloudflare' SDK client.