(fix) test: Fix KeyError in test_idle_heartbeat with connection replacement

The test_idle_heartbeat test was failing with a KeyError when connections were dynamically replaced during the test run in shard-aware environments.
Root cause: The test captures connection object IDs at the start, sleeps for heartbeat intervals, then tries to validate those connections.
In shard-aware ScyllaDB deployments, the driver may replace connections during this window.
When a connection is replaced, the new connection object has a different Python ID, causing a KeyError on lookup.

Fix: Skip validation for connections that weren't present in the initial snapshot.
This preserves the test's intent (validating heartbeats on stable connections) while being robust to dynamic connection management in shard-aware mode.

I'm not sure this is a great fix. I don't think it has anything to do with this branch. I'll push it and if it's OK, I'll also create a separate PR for it.
Why it popped now?

Signed-off-by: Yaniv Kaul <yaniv.kaul@scylladb.com>

Author: mykaul

tests/integration/standard/test_cluster.py

@@ -769,8 +769,14 @@ def test_idle_heartbeat(self):
         connections = [c for holders in cluster.get_connection_holders() for c in holders.get_connections()]
 
         # make sure requests were sent on all connections
+        # Note: connections can be replaced in shard-aware environments, so we skip
+        # connections that weren't present in the initial snapshot
         for c in connections:
-            expected_ids = connection_request_ids[id(c)]
+            conn_id = id(c)
+            if conn_id not in connection_request_ids:
+                # Connection was replaced during the test, skip validation
+                continue
+            expected_ids = connection_request_ids[conn_id]
             expected_ids.rotate(-1)
             with c.lock:
                 assertListEqual(list(c.request_ids), list(expected_ids))