A few reliability fixes, and clearer errors when screenshots are uploaded to your storage.
Fewer failed requests
- Busy rendering servers. During traffic spikes, a rendering server could fail to open a page in time, and the request failed with an internal error. It’s now retried on another server.
- Pages that redirect themselves. When you request HTML or markdown, a page that redirected or reloaded itself right after loading could fail with an internal error. You now get the content of the page it ends up on.
- Request logging. Executed requests are logged more reliably during traffic spikes.
Clearer storage errors
When a screenshot is uploaded to your storage, configuration errors now say what to fix, for example, an access key ID the storage doesn’t know, an endpoint that already includes the bucket name, a bucket that doesn’t exist, or a storage class your storage doesn’t support. For Cloudflare R2, the errors also point to the key to use.
