since my worker is on a custom domain, could I just use a Cache rule for its responses?

Cache can cache any Request/Response within your Worker, but your Worker will always be invoked, even if you serve a cached response

Hey @HardAtWork, thanks for responding

I get that, but what would be the optimal way to cache the worker's actual response, if any?

I am not running any calls inside the worker

It's a GET call with a JSON response

You cannot. If a Worker is on a Route, it will be called, whether a Response is cached or not

ah I see

so the cache api is only for in-worker request/response

good thing I asked before venturing into nothingness 🙂

thanks again

Has anyone gotten ffmpeg.wasm to work on workers? I can't get it to load

Currently there is no possibility to access exception "logs" of workers after they occurred, right? I have a customer that experiences an error that I can't reproduce because of external APIs I don't have access to in production. Suddenly errors arise which don't occur with my test instance but I can't sit the whole day in front of the realtime log and wait for the customer to execute the action. I see that errors are occurring + customer sent me screenshot of the default workers error page (helpful 😛) but I don't see any way to access the error. Just a message would be helpful already

I read something about logpush but that sounds complicated to setup just for this simple use-case

If you don't want to use logpush, something like sentry is what i'd say

Log engine is to allow that but for now you have to use logpush and 3rd party service. Luckily logpush for workers isnt ent only 🙂

Logpush or use something like sentry in the code 🤷

D1 can be queried from an Email Worker now, if you install

wrangler@beta

, cc @Skye @Chaika

Nice, good to hear that something is in the works :)

Nothing to do in particular with my first question but do you guys use Sentry in prod workers? It sounds like quite a bit of overhead for applications that don't run 24/7 (e.g. workers/lambda/whatever)

And if anyone cared to measure, does it impact CPU time?

Not that I'd be concerned but just because I am wondering

thanks!

so what I do, is have logpush exporting to IDrive E2 and Wasabi. Both don't charge for API write requests, so I'm only paying a few pennies per gigabyte.

you could push to AWS S3, and then have a trigger to basically grep through the files.

has anyone that used web crypto ran into this problem:

Error: DataError, Invalid PKCS8 input., 0

Why is VS Code complaining about this function not existing?

where is it getting the types from