The CloudSQL proxy pod was using backstage.selectorLabels which made it
match the Backstage service selector (app.kubernetes.io/name: backstage,
app.kubernetes.io/instance: backstage). This caused Traefik to sometimes
route requests to the CloudSQL proxy instead of Backstage, resulting in
502 errors.
Changed pod labels to use distinct values (backstage-cloudsql-proxy)
so the pod is not selected by the Backstage service.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
