Hi Railway team,

I've got a Pre-deploy Command that isn't fully running, and I've dug in enough to be pretty confident it's not a mistake on my end —

hoping you can take a look.

Project: handsome-courtesy (8ed175ba-9951-48f5-833a-caf781974ac0)

Service: livoh

Environment: production (8c584d9d-4135-4092-a964-127113a12cb6)

Configured Pre-deploy Command:

npx prisma migrate deploy && node dist/db/loadStoredProcedures.js

What's happening: prisma migrate deploy runs and completes every time — I can see its full output in Deploy Logs. But node

dist/db/loadStoredProcedures.js, the second half of that same &&-chained command, never produces a single line of output. Not

success, not a no-op message, not even an error — it's designed to log loudly on every code path (including a hard failure if an env

var is missing), so total silence points to it not running at all rather than running and failing quietly.

I've reproduced this across three separate deployments so far: b1c23194-d1b3-41af-8f08-c140332d2b55,

79b9e66b-be2f-4ae8-bf0c-f6c7a8018049, and b1423715-d381-4de7-98e9-d86d38d2c33a. In each case, Deploy Logs go straight from the Prisma

migration output to the app's own runtime startup logs, with nothing in between.

To rule out a mistake on my side, I've already:

• Confirmed the command itself is correct (it matches my package.json's own start script, split across your Pre-deploy/Start fields).
• Re-saved the Pre-deploy Command field and triggered a fresh deploy from the dashboard — same result.
• Checked Deploy Logs directly in the dashboard, not just the CLI, and confirmed the same absence there.
• Confirmed the identical script runs and logs correctly ("loadStoredProcedures: script applied" or "unchanged — skipped") on every

boot of our staging environment, which is configured the same way. It's only production where the second command seems to be dropped.

• Confirmed the script itself works fine when I run it manually against the production database from my own machine.

So the app code and database are fine — this is specifically about the Pre-deploy step not carrying through its second command on

this one environment.

Could someone take a look at why production's Pre-deploy Command isn't executing that second command? Happy to provide any further

logs or details you need.

Thanks,

Kartikey