I am either crazy... or I found a subtle/silent way to break queue producing
Hello. I will do more digging into this as I am able... but wanted to share in case someone can provide direction for how to gather more details:
- I have a worker that consumes from Queue A and produces into Queue B
- That worker reaches out to supabase using the supabase-js library
- After the supabase upsert/select... the
env.QUEUE_B.send() method ends the invocation
I can env.QUEUE_B.send() multiple times up until that point... but once I supabase.schema().upsert().select().single() (which returns without error), the next env.QUEUE_B.send() ends the program with no errors, no warnings, nothing. try/catch gives nothing ,etc.
This sounds nuts, but it seems like there is some weird / latent global state that supabase or some wrapper code is modifying that changes behavior.
Is there any way to gather more info on this type of thing and what is happening? I had to get something shipped, so I switched to fetching against the other worker directly rather than the async / queueing approach.0 Replies