Add metric to report about sync standby replica status (#2615)

Close #2613

Co-authored-by: Alexander Kukushkin <[email protected]>

docs/rest_api.rst

@@ -139,6 +139,9 @@ Retrieve the Patroni metrics in Prometheus format through the ``GET /metrics`` e
 	# HELP patroni_replica Value is 1 if this node is a replica, 0 otherwise.
 	# TYPE patroni_replica gauge
 	patroni_replica{scope="batman"} 0
+	# HELP patroni_sync_standby Value is 1 if this node is a sync standby replica, 0 otherwise.
+        # TYPE patroni_sync_standby gauge
+        patroni_sync_standby{scope="batman"} 0
 	# HELP patroni_xlog_received_location Current location of the received Postgres transaction log, 0 if this node is not a replica.
 	# TYPE patroni_xlog_received_location counter
 	patroni_xlog_received_location{scope="batman"} 0

patroni/api.py

@@ -269,6 +269,10 @@ class RestApiHandler(BaseHTTPRequestHandler):
         metrics.append("# TYPE patroni_replica gauge")
         metrics.append("patroni_replica{0} {1}".format(scope_label, int(postgres['role'] == 'replica')))
 
+        metrics.append("# HELP patroni_sync_standby Value is 1 if this node is a sync standby replica, 0 otherwise.")
+        metrics.append("# TYPE patroni_sync_standby gauge")
+        metrics.append("patroni_sync_standby{0} {1}".format(scope_label, int(postgres.get('sync_standby', False))))
+
         metrics.append("# HELP patroni_xlog_received_location Current location of the received"
                        " Postgres transaction log, 0 if this node is not a replica.")
         metrics.append("# TYPE patroni_xlog_received_location counter")
@@ -686,6 +690,10 @@ class RestApiHandler(BaseHTTPRequestHandler):
             if result['role'] == 'replica' and self.server.patroni.ha.is_standby_cluster():
                 result['role'] = postgresql.role
 
+            if result['role'] == 'replica' and cluster and cluster.is_synchronous_mode()\
+                    and cluster.sync and postgresql.name in cluster.sync.members:
+                result['sync_standby'] = True
+
             if row[1] > 0:
                 result['timeline'] = row[1]
             else:

tests/test_api.py

@@ -184,6 +184,7 @@ class TestRestApiHandler(unittest.TestCase):
 
     def test_do_GET(self):
         MockPatroni.dcs.cluster.last_lsn = 20
+        MockPatroni.dcs.cluster.sync.members = [MockPostgresql.name]
         MockRestApiServer(RestApiHandler, 'GET /replica')
         MockRestApiServer(RestApiHandler, 'GET /replica?lag=1M')
         MockRestApiServer(RestApiHandler, 'GET /replica?lag=10MB')
@@ -196,7 +197,6 @@ class TestRestApiHandler(unittest.TestCase):
         with patch.object(RestApiHandler, 'get_postgresql_status', Mock(return_value={'state': 'running'})):
             MockRestApiServer(RestApiHandler, 'GET /health')
         MockRestApiServer(RestApiHandler, 'GET /leader')
-        MockPatroni.dcs.cluster.sync.members = [MockPostgresql.name]
         MockPatroni.dcs.cluster.is_synchronous_mode = Mock(return_value=True)
         with patch.object(RestApiHandler, 'get_postgresql_status', Mock(return_value={'role': 'replica'})):
             MockRestApiServer(RestApiHandler, 'GET /synchronous')