Investigate behaviour if file-server times out at startup of instance

### Description

Follow-up task to https://github.com/flowforge/CloudProject/issues/184

If a request to the file-server service times out when NR is starting, it appears to leave the instance in an odd state - and requires a full suspend/resume to recover.

This task is to investigate that scenario and to ensure it is automatically recoverable.

### Epic/Story

_No response_

Start by tracing the Node-RED startup path that requests the file-server service, then reproduce a timeout during instance startup. Check how the instance enters its odd state and what suspend/resume currently resets. Done means a file-server timeout leaves the instance recoverable automatically, with coverage for the timeout scenario.

Written by the indexing model from the issue text.