fix(tests): increase Bootstrap Talos cluster timeout to 60s

On fast ephemeral runners a freshly booted Talos node has not yet
converged NTP when the previous 10s outer timeout fired, so the
embedded 'until talosctl bootstrap; do sleep 1; done' retry loop only
got one 5-6s attempt before the loop was killed. Every RPC returned
'FailedPrecondition: time is not in sync yet' and E2E failed in the
Prepare environment step.

Bump the outer timeout to 60s and the inter-attempt sleep to 2s. This
aligns with the surrounding bootstrap gates (etcd members check at
180s, drain RPC errors at 60s, node registration at 60s) and gives
NTP plenty of time to converge on ephemeral runners.

Assisted-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>

hack/e2e-prepare-cluster.bats

@@ -235,8 +235,11 @@ EOF
 }
 
 @test "Bootstrap Talos cluster" {
-  # Bootstrap etcd on the first node
-  timeout 10 sh -ec 'until talosctl bootstrap -n 192.168.123.11 -e 192.168.123.11; do sleep 1; done'
+  # Bootstrap etcd on the first node.
+  # Retry for up to 60s: talosctl bootstrap refuses with "time is not in
+  # sync yet" until NTP converges on a freshly booted node, which on fast
+  # ephemeral runners can take several attempts of ~5s each.
+  timeout 60 sh -ec 'until talosctl bootstrap -n 192.168.123.11 -e 192.168.123.11; do sleep 2; done'
 
   # Wait until etcd is healthy
   if ! timeout 180 sh -ec 'until talosctl etcd members -n 192.168.123.11,192.168.123.12,192.168.123.13 -e 192.168.123.10 >/dev/null 2>&1; do sleep 1; done'; then