Revert --token-ttl=0 on dashboard - broke login entirely

Login stopped working the moment this shipped. Suspect this dashboard
version treats token-ttl=0 as "expire immediately" rather than "never
expire". Reverting to the default (900s) to restore working login;
the idle-timeout annoyance can be revisited with a large finite value
instead of 0, tested before shipping.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

manifests/alaska/k8s-dashboard.yaml

@@ -62,13 +62,12 @@ spec:
         args:
           - --auto-generate-certificates
           - --namespace=kubernetes-dashboard
-          # Disables the default 15-minute idle session timeout, so a
+          # --token-ttl=0 was tried here to disable the 15-min idle
-          # logged-in session persists instead of dropping back to the
+          # timeout, but login broke immediately after that pod came up -
-          # token screen. The service account token itself already
+          # in this dashboard version, 0 appears to mean "expire
-          # doesn't expire (it's a Secret-backed legacy token, not a
+          # immediately" rather than "never expire". Reverted; back to
-          # short-lived TokenRequest one) - this was purely the
+          # the default 900s timeout until a properly-tested value (e.g.
-          # dashboard's own session cache.
+          # a long finite number of seconds) is confirmed safe.
-          - --token-ttl=0
         volumeMounts:
         - name: kubernetes-dashboard-certs
           mountPath: /certs