Stop failing on unknown ceilometer metrics

The metric test is too strict and fails if it finds a metric that
is not explicitly listed in the test. Given the test does check
that the expected metrics are present there seems not need to fail
if a new one has added upstream.

zaza/openstack/charm_tests/ceilometer_agent/tests.py

@@ -92,7 +92,7 @@ class CeilometerAgentTest(test_utils.OpenStackBaseTest):
         unexpected_found_metric_names = (
             found_metric_names - expected_metric_names)
         if len(unexpected_found_metric_names) > 0:
-            self.fail(
+            logging.info(
                 'Unexpected metrics '
                 'published: ' + ', '.join(unexpected_found_metric_names))