I temporarily removed the predeploy script and the service is now deploying correctly 👀

What type of script are you running on pre-deploy? migrations?

Mind sharing that service link?

Here you go:

https://railway.com/project/b7ffa58d-4fc3-417f-8ad8-ffc29078d847/service/da0c7c88-cc65-43f8-85ef-d3523f41d186?groupId=525b88bb-0043-43cf-ad4d-5b1bf38ff530&environmentId=e29cbcaf-c244-4fb9-a607-934f68f47aca

Yeah, migrations

The migrations complete though, and then it just hangs

(the "stopping container" only happened after I aborted the deploy)

🤔

Any recent changes to the pre-deploy command? Was it running fine before?

No changes, I don't even have new migrations 😅

Yeah been running fine for months

!t

I'm not seeing anything that is hanging, could I get a link to that?

This deploy: https://railway.com/project/b7ffa58d-4fc3-417f-8ad8-ffc29078d847/service/da0c7c88-cc65-43f8-85ef-d3523f41d186?groupId=525b88bb-0043-43cf-ad4d-5b1bf38ff530&environmentId=e29cbcaf-c244-4fb9-a607-934f68f47aca&id=365a9806-28ef-468c-88f4-507a666a5bcd

It sat for 20 minutes before I removed the predeploy command and redeployed

Deploying again with predeploy

For context, this is the script, so if the Done Migrations line is hit, there is nothing more

It isn't a big deal for me, I'm just messing around on my staging for now, but I just wanted to raise it incase it is a broader issue 🙂

Oh

That worked

You made the change that fast?

I did

28 second build too, nice

Undoing the os.Exit to double check that was it

Yeah it hangs again. That is new behaviour then, no? I haven't updated this script in a while

I see two other reports, so something is up on our end.

Ah alright, good luck! <:salute:1137099685417451530>

Thanks for the help as always <:bufo_pray:1331539894769684513>

No problem!

sorry for the "same", but same

As suggested previously, try adding an exit command at the end of your script!

i assume the previous behavior will be restored? would rather wait then, my predeploy cmd is currently just running drizzle-kit migrate. would like to avoid wrapping it in a custom script on 3 different services

drizzle-kit migrate definitely exits with 0 on success btw 🤔

I haven't attempted to reproduce yet, and even so, it's unfortunately not something we will tackle over the weekend.

Could you maybe clarify what this new requirement is for predeploy commands? I understand in the other case adding an explicit exit made it work, but my command already exits with 0 (and always has), so I don't really see what else I can do at the moment to make it work.

Also, ofc understand it's a weekend, but if this new behavior is unintended and unclear, feels like this would warrant an incident? 🤔

It's not a new requirement, but I have yet to confirm if it is a regression. An incident won't be called, as incidents are for platform-wide issues that affect the majority of the platform.

Alrighty. Re it being a regression - I can say in our case no changes have been made to anything pre-deploy related since our last successful deploy for this same service on Friday at Nov 28, 2025, 3:58:37 PM GMT+1. Today since (at the latest) Nov 29, 2025, 5:26:24 PM GMT+1 our deploys are stuck.

@json - Could I ask you to open your own thread? it makes it easier for us to track.