Unless theres data you cant afford to loose I suggest deleting the faulty service and creating a new one then referencing the new one in your code

What was misconfigured ?
Can you remove the railway.toml file ? If this is coming from a repo then you need to delete it from your repo.

Tnx for the responses
So my project is a repo based project and the posgres DB was created separately into the project .
And now somehow the posgres part has changed and is also railway.toml based . And i can not edit this part , to change it back.
Deleting the .toml file is not a good ideea because the vendure backend is based on that .

That’s definitely odd behavior. In my experience when I deploy a Postgres DB on Railway, it just uses the managed image from ghcr.io, so it shouldn’t be affected by `railway.toml`. Is there any chance the MCP you added also spun up its own Postgres instance? It might explain why things are suddenly being treated differently — maybe there are two DBs now and one got pulled into your service config by mistake?

Could you share your railway.toml file and a screenshot of your Railway services page?

Finally i deleted and created again . like sim wrote yes the Postgres DB on Railway, it just uses the managed image from ghcr.io,
but if this somehow changes you can not switch back.

here is an other issue

i have this in my railway.toml
[service.vendure-worker.deploy]

startCommand = "node dist/index-worker.js"

restartDelay = 5

but is not taking any effect. still the index.js is deployed

https://docs.railway.com/reference/build-and-start-commands#start-command
Might give a bit of an answer on that.