Who can tell me why the Worker runtime environment of Cloudflare sometimes gets blocked for 15 minutes?
It seems that this request has already died, but why is it still blocking? It would be fine if there were some mistakes in the report. But now it just keeps stopping and doesn't move any further.
If there is a panic in your worker, then the request will stall.
We are working on a fix here to turn panics into immediate aborts which I am working on releasing in the coming release this week.
Is there any progress? This is a problem I often encounter.
The first fix we landed is that the Wasm is reinitialized automatically if running workers-rs after 0.6.7. This means the stalled requests will just be a few requests when there is a panic. The panic fix is ongoing work in wasm bindgen. In the latest version of wasm bindgen we now support building with unwind support in Rust. Follow-on to enable error recovery is in progress. Watching the wasm bindgen releases upstream for exception handling updates is one way to follow this work.