This issue is coming from a comment at https://crrev.com/c/872514#message-6f6accffbbdc6f81d639b47a892f4c5ac101926e:
---
Note that while foreign fetch was the only feature using this custom timeout thing, it is probably needed for some regular fetches as well:

if ever nested workers are introduced, fetches for/from workers that are owned by the SW should probably not give the SW more time to execute
fetches or icons etc for notifications initiated by a SW (which should be intercepted by the SW according to https://github.com/w3c/ServiceWorker/issues/1161 and some F2F discussions I remember in the past, although I can't find exactly where that is specified) should probably not let the SW extend its timeout either.
And there might be more, those are just the first two that come to mind.

---

We may need to dig into more about how we deal with the timeout for fetch events.
Has there already been discussion about that?