chore: improve monitoring for fix: gracefully handle transient 502 errors during

Added `ignoreErrors` to the Sentry `init()` configuration in `SentryInitEffect.tsx` to filter out transient browser network errors that are not actionable:

- `NetworkError when attempting to fetch resource` — Firefox's TypeError for failed fetch
- `Failed to fetch` — Chrome/Edge's TypeError for failed fetch
- `Load failed` — Safari's TypeError for failed fetch

These errors occur when the browser loses connection to the server (e.g., during a transient 502 from Cloudflare). They are not bugs — they are symptoms of transient infrastructure availability that self-resolve. Capturing them in Sentry creates noise without actionable signal.

packages/twenty-front/src/modules/error-handler/components/SentryInitEffect.tsx

@@ -47,6 +47,12 @@ export const SentryInitEffect = () => {
                 onunhandledrejection: false, // handled in PromiseRejectionEffect
               }),
             ],
+            ignoreErrors: [
+              // Transient browser network errors from reverse proxy 502/503/504
+              'NetworkError when attempting to fetch resource',
+              'Failed to fetch',
+              'Load failed',
+            ],
             tracePropagationTargets: [
               'localhost:3001',
               REACT_APP_SERVER_BASE_URL,