tests: fix flaky SSL test by increasing connect timeout and retry budget

The routes_visible() polling function in TestSslThroughNlb creates a new
TestCluster with SSL on every retry attempt.  Under resource pressure
(--smp 2 --memory 2048M shared across 3 nodes), the SSL handshake plus
CQL negotiation can exceed the default 5-second connect_timeout, causing
intermittent OperationTimedOut failures.

Fix by passing connect_timeout=30 to TestCluster (matching the generous
timeout recommended for slow-starting clusters) and increasing the
wait_until_not_raised parameters from (0.5, 10) to (1, 30), consistent
with other wait_until_not_raised calls in this file (lines 773, 855).

Author: mykaul

tests/integration/standard/test_client_routes.py

@@ -1059,7 +1059,7 @@ def test_ssl_without_hostname_verification_through_nlb(self):
             def routes_visible():
                 with TestCluster(
                     contact_points=["127.0.0.1"],
-                    ssl_context=ssl_ctx,
+                    ssl_context=ssl_ctx, connect_timeout=30,
                 ) as c:
                     session = c.connect()
                     rs = session.execute(
@@ -1071,7 +1071,7 @@ def routes_visible():
 
             wait_until_not_raised(
                 lambda: self.assertTrue(routes_visible()),
-                0.5, 10,
+                1, 30,
             )
 
             with Cluster(