[Ekhorn]

Description

Relates to #3526

The manager could stay on a blank page when a request failed during startup, for example while the backend restarts. It stayed blank after the backend was back, until a reload:

• The app config fetch had no catch, so on a network error OrApp was never added to the page.
• loadManagerInfo parsed the response inside the XHR load listener. An HTML error page, such as a proxy's 503, threw there, so Manager.init never settled.

A backend restart triggers this more often than expected. The UI files only carry Last-Modified, so a new tab loads index.html and the bundle from the browser cache and starts the app while the API is still down.

Both cases were reproduced in Chromium on a staging server running main, by failing these requests in the browser.

Reproduction

1. In Chrome DevTools, open Request conditions, enable blocking and throttling, and add the full URL https://<host>/api/master/configuration/manager.
2. Reload. Without this PR the page stays blank, and the console only shows Failed to fetch. With it, the login page loads.

Changelog

• Fall back to the default app config when the manager app's config fetch fails
• Reject in loadManagerInfo when the info response isn't JSON, so initialization continues
• Add app tests that fail each request and expect the login page

Checklist

• 1. Acceptance criteria of the linked issue(s) are met
• 2. Tests are written and all tests pass
• 3. Changes are manually tested by you and the reviewer
• 4. Documentation is written or updated

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[Ekhorn]

We discussed we want to show a loading spinner after an initial request failure or when it takes several seconds to load. If it fails we want exponential backoff and after ~30-60s we want to show something like "could not load app, try again later".