I'm currently evaluating Railway as an alternative to Render for our project. While most of the migration has been successful, I'm facing an issue with my pre-deploy command that appears to be an Out of Memory (OOM) error, but I'm not receiving adequate feedback to troubleshoot.

Configuration

My preDeployCommand is set to:

pnpm run --filter @repo/backend pg:migrate

This executes the following command in my apps/backend directory:

NODE_OPTIONS='--import tsx' drizzle-kit migrate

Error Details

When the deploy process runs, I see the following in my Deploy Logs tab:

Starting Container

> fotw-admin@ pg:migrate /app
> pnpm run --filter @repo/backend pg:migrate


> @repo/backend@1.0.0 pg:migrate /app/apps/backend
> NODE_OPTIONS='--import tsx' drizzle-kit migrate

No config path provided, using default 'drizzle.config.ts'
Reading config file '/app/apps/backend/drizzle.config.ts'
Killed

/app/apps/backend:
 ERR_PNPM_RECURSIVE_RUN_FIRST_FAIL  @repo/backend@1.0.0 pg:migrate: `NODE_OPTIONS='--import tsx' drizzle-kit migrate`
Exit status 137
 ELIFECYCLE  Command failed with exit code 137.

Stopping Container

The exit code 137 typically indicates an Out of Memory (OOM) error. The command does run successfully on my local computer and in my Render environment.

Questions

1. How can I further debug this issue to determine the exact memory requirements?

2. Is it possible to SSH into the container running the pre-deploy command for direct troubleshooting?

3. Are there any metrics available for the pre-deploy command container? I don't see any in my metrics tab.