For clients that use sprockets.mixins.metrics,

DB metrics use a "db" subpath.

For example, if a program calls self.postgres_execute() with
metric_name='insert-user', that will record a metric on path
"db.insert-user".

This is particularly beneficial for visualizing metrics.
By querying Graphite for all metrics that have "db" in their path,
you can easily visualize all DB metrics for a program.
If someone later adds another DB query and metric to that program,
it meets the same conditions,
thus it can be automatically added to the graph.

sprockets_postgres.py

@@ -796,7 +796,7 @@ class RequestHandlerMixin:
         if hasattr(self, 'influxdb'):  # sprockets-influxdb
             self.influxdb.set_field(metric_name, duration)
         elif hasattr(self, 'record_timing'):  # sprockets.mixins.metrics
-            self.record_timing(duration, metric_name)
+            self.record_timing(duration, 'db', metric_name)
         else:
             LOGGER.debug('Postgres query %s duration: %s',
                          metric_name, duration)

tests.py

@@ -444,8 +444,9 @@ class RequestHandlerMixinTestCase(TestCase):
         self.assertEqual(response.code, 200)
         self.assertEqual(json.loads(response.body)['value'], expectation)
         self.app.record_timing.assert_called_once()
-        duration, metric_name = self.app.record_timing.call_args[0]
+        duration, metric_path, metric_name = self.app.record_timing.call_args[0]
         self.assertGreater(duration, 0.0)
+        self.assertEqual('db', metric_path)
         self.assertEqual('', metric_name)
 
     def test_postgres_multirow_get(self):