dt/audit: Add 'rpc::service_unavailable' to upgrade test allow list

This test works by performing a rolling upgrade and verifying audit log
function at each stage. With RPC-based audit logging enabled, the audit
client on some node can race against service startup on some other, recently
upgraded node, resulting in transient RPC error logs on the source node.
In these cases the audit log client will immediately retry to avoid disruption,
but the error log itself will trigger BadLogLines.

Since the test is working as intended and there is no expectation of
synchronization between RPC client and service on separate nodes, let's
just add the specific rpc error log to the allow list for this test.

Fixes CORE-15036

Signed-off-by: Oren Leiman <oren.leiman@redpanda.com>

Author: oleiman

tests/rptest/tests/audit_log_test.py

@@ -3374,7 +3374,11 @@ def stop_cond(records):
         return len(self.read_all_from_audit_log(filter_fn, stop_cond))
 
     @skip_fips_mode
-    @cluster(num_nodes=5, log_allow_list=RESTART_LOG_ALLOW_LIST)
+    @cluster(
+        num_nodes=5,
+        log_allow_list=RESTART_LOG_ALLOW_LIST
+        + [re.compile("kafka/data/rpc - .*rpc::errc::service_unavailable.*")],
+    )
     def test_audit_log_upgrade_all_nodes(self):
         """
         Test that audit logging works on all nodes during rolling upgrade