can you share logs, tech stack...

Hi — I think both issues (cron reliability and PostgreSQL not working) are likely related to how the services are configured and running on Railway rather than pure code problems.

From experience, most Railway issues come down to service lifecycle behavior, environment configuration, or container restarts.

Here’s how I’d systematically debug and fix this: PostgreSQL Not Connecting The database is usually created correctly.

The issue is almost always in the connection setup.

Please check the following: Make sure the environment variables are added to the correct backend service.

Confirm that DATABASE_URL is defined inside the actual backend service, not just at the project level or another service.

Verify the variable name.

If your app expects DATABASE_URL but you set DB_URL or POSTGRES_URL, it won’t connect.

Use the internal/private Postgres URL if both backend and Postgres are on Railway.

Avoid using the public connection string unless necessary.

Redeploy or restart the backend after adding environment variables.

Containers may not pick up new variables until redeployed.

Check the exact error in logs: “password authentication failed” usually means wrong credentials.

“ENOTFOUND” or “host not found” means incorrect host or wrong URL type.

SSL errors may require adding ?sslmode=require depending on the client.

If you’re using an ORM, run migrations.

A database existing does not mean tables are created.

Prisma: npx prisma migrate deploy Sequelize: sequelize db:migrate Cron Jobs Not Running Reliably If your cron jobs are running inside the same web service using node-cron or setInterval, this is very likely a container lifecycle issue.

On Railway: If Serverless or app sleeping is enabled, the service may pause when there’s no traffic.

If the service restarts due to memory limits, deploys, or crashes, in-memory cron jobs stop.

If multiple replicas are running, jobs may duplicate or behave inconsistently.

Please verify: Whether Serverless is enabled. How many replicas are running.

If deployment logs show restarts or cold starts. The restart policy configuration.

For production reliability, cron jobs should ideally run in a separate worker service or use Scheduled Executions instead of staying inside the main web service.

Why both problems might be connected If the service is restarting: Database connections drop.

Cron jobs stop.

Background simulations halt.

External API connections fail without recovery.

I would first review: Deployment logs.

Container restart events.

Memory usage.

Replica configuration.

Environment variable setup.

If you can share: The exact database connection error from logs.

Whether cron runs inside the web service or a separate worker.

Your replica configuration.

I can help pinpoint the root cause quickly.

Try connecting to your DB via psql or datagrip (or any other rdbms) to see if you are using the correct db url. If you go into your Postgresql service -> Variables
there will be 2 urls: DATABASE_PUBLIC_URL and just DATABASE_URL, which one do you use? Did you specify the correct database name (the last word after "/" symbol) ?