It’s what you are describing, and it’s not a client problem. It’s impossible for a client to solve. You can’t tell whether a timed out request succeeded or not.

Idempotent network requests are a standard feature of many APIs - sending the same request multiple times should result in only one action being performed, but Lemmy doesn’t support them yet.

So yes, it’s “caused” by retries, but the bug is that the backend doesn’t properly support retries. Clients don’t do anything wrong.