Queues: Broken locally when using multip...
Any plans on supporting queues started with multiple dev commands?
(https://github.com/cloudflare/workers-sdk/issues/9795)
GitHub
Queues: Broken locally when using multiple dev commands · Issue #9...
What versions & operating system are you using? MacOS latest, node 24, wrangler 4.22.0 Please provide a link to a minimal reproduction https://github.com/astanciu/cf_worker_repro Describe the B...
5 Replies
It is in the backlog but we don't have it lined up for implementation very soon. We discussed the idea of storing persistence for all the Workers in the dev registry in the same place (a bit like them sharing resources in a CF account).
In this case a queue could be stored in a similarly central location.
don't have it lined up for implementation very soonSad to hear 😦 In order to match production, to some degreen, we currently implementing this workaround: - producer puts a message on a queue (local queue, that only the producer can consume) - producer consumes their own queue, and on msg sends the msg to via RPC to the target worker - consumer receives message via RPC, then does a send() on their local queue, this way the actual queue() handler gets properly invoked. I don't get how more of your customers using queues aren't having this issue... sending messages between two distinct workers seems like such a basic, first class, feature 😦
https://developers.cloudflare.com/queues/configuration/local-development/#separating-producer--consumer-workers
Can't you start them both at the same time and produce / consume from different workers?
no, that doesn't work for a number of reasons, biggest of which are:
- workers started with -c -c method aren't individually addressable, the only url you get is the one for the "main" worker
- simply does not work for workers using Vite
Ahh gotcha