Dear Cafe, this topic has been discussed [1,2] a decade ago. What is the current state of affairs in this matter? It still seems to be addressed neither in yesod nor warp [*]. We have two use cases: (1) A shutdown route that allows admins to cleanly shut down the web application. (2) We want certain exceptions to terminate the application and have the surrounding monitoring (systemd, docker, ...) re-start the application. For (2) we are safe, I think, because the worker threads are forked from the main thread [3] calling warp, so propagating the exceptions up to the top bypasses warp and yesod's very comprehensive exception handlers. But any exception thrown while answering a request (such as (1)) is caught by default. Thanks Olaf [1] https://groups.google.com/g/yesodweb/c/VoenrabRUBQ [2] https://stackoverflow.com/questions/7881327/how-do-i-implement-a-shutdown-co... [3] https://mail.haskell.org/pipermail/haskell-cafe/2022-March/135132.html [*] Warp offers setOnException and setOnExceptionResponse. But does re- throwing certain exceptions from there terminate warp?