Fix deployment count metric

kubernetes_state/datadog_checks/kubernetes_state/kubernetes_state.py

@@ -94,7 +94,6 @@ def __init__(self, name, init_config, instances):
                 'allowed_labels': ['storageclass', 'phase'],
             },
             'kube_service_spec_type': {'metric_name': 'service.count', 'allowed_labels': ['namespace', 'type']},
-            # is a count by namespace and phase <Active|Terminating>
             'kube_namespace_status_phase': {'metric_name': 'namespace.count', 'allowed_labels': ['phase']},
             'kube_replicaset_owner': {
                 'metric_name': 'replicaset.count',
@@ -125,12 +124,11 @@ def __init__(self, name, init_config, instances):
             'kube_node_spec_unschedulable': self.kube_node_spec_unschedulable,
             'kube_resourcequota': self.kube_resourcequota,
             'kube_limitrange': self.kube_limitrange,
-            'kube_persistentvolume_status_phase': self.count_objects_by_tags,
+            'kube_persistentvolume_status_phase': self.sum_values_by_tags,
             'kube_service_spec_type': self.count_objects_by_tags,
-            'kube_namespace_status_phase': self.count_objects_by_tags,
+            'kube_namespace_status_phase': self.sum_values_by_tags,
             'kube_replicaset_owner': self.count_objects_by_tags,
             'kube_job_owner': self.count_objects_by_tags,
-            # to get overall count is to filter by Available
             'kube_deployment_status_observed_generation': self.count_objects_by_tags,
         }
 
@@ -869,6 +867,25 @@ def kube_limitrange(self, metric, scraper_config):
         else:
             self.log.error("Metric type %s unsupported for metric %s", metric.type, metric.name)
 
+    def sum_values_by_tags(self, metric, scraper_config):
+        """ Sum values by allowed tags and submit counts as gauges. """
+        config = self.object_count_params[metric.name]
+        metric_name = "{}.{}".format(scraper_config['namespace'], config['metric_name'])
+        object_counter = Counter()
+
+        for sample in metric.samples:
+            tags = []
+            for l in config['allowed_labels']:
+                tag = self._label_to_tag(l, sample[self.SAMPLE_LABELS], scraper_config)
+                if tag is None:
+                    tag = self._format_tag(l, "unknown", scraper_config)
+                tags.append(tag)
+            tags += scraper_config['custom_tags']
+            object_counter[tuple(sorted(tags))] += sample[self.SAMPLE_VALUE]
+
+        for tags, count in iteritems(object_counter):
+            self.gauge(metric_name, count, tags=list(tags))
+
     def count_objects_by_tags(self, metric, scraper_config):
         """ Count objects by allowed tags and submit counts as gauges. """
         config = self.object_count_params[metric.name]
@@ -883,7 +900,7 @@ def count_objects_by_tags(self, metric, scraper_config):
                     tag = self._format_tag(l, "unknown", scraper_config)
                 tags.append(tag)
             tags += scraper_config['custom_tags']
-            object_counter[tuple(sorted(tags))] += sample[self.SAMPLE_VALUE]
+            object_counter[tuple(sorted(tags))] += 1
 
         for tags, count in iteritems(object_counter):
             self.gauge(metric_name, count, tags=list(tags))

kubernetes_state/tests/fixtures/prometheus.txt

@@ -100,9 +100,9 @@ kube_deployment_spec_strategy_rollingupdate_max_unavailable{deployment="tiller-d
 # HELP kube_deployment_status_observed_generation The generation observed by the deployment controller.
 # TYPE kube_deployment_status_observed_generation gauge
 kube_deployment_status_observed_generation{deployment="failingtest",namespace="default"} 1
-kube_deployment_status_observed_generation{deployment="jaundiced-numbat-kube-state-metrics",namespace="default"} 1
+kube_deployment_status_observed_generation{deployment="jaundiced-numbat-kube-state-metrics",namespace="default"} 4
 kube_deployment_status_observed_generation{deployment="kube-dns",namespace="kube-system"} 1
-kube_deployment_status_observed_generation{deployment="tiller-deploy",namespace="kube-system"} 1
+kube_deployment_status_observed_generation{deployment="tiller-deploy",namespace="kube-system"} 5
 # HELP kube_deployment_status_replicas The number of replicas per deployment.
 # TYPE kube_deployment_status_replicas gauge
 kube_deployment_status_replicas{deployment="failingtest",namespace="default"} 1

kubernetes_state/tests/test_kubernetes_state.py

@@ -74,7 +74,6 @@
     NAMESPACE + '.container.gpu.request',
     NAMESPACE + '.container.gpu.limit',
     # replicasets
-    NAMESPACE + '.replicaset.count',
     NAMESPACE + '.replicaset.replicas',
     NAMESPACE + '.replicaset.fully_labeled_replicas',
     NAMESPACE + '.replicaset.replicas_ready',
@@ -100,10 +99,7 @@
     NAMESPACE + '.resourcequota.limits.memory.limit',
     # limitrange
     NAMESPACE + '.limitrange.cpu.default_request',
-    # services
-    NAMESPACE + '.service.count',
     # jobs
-    NAMESPACE + '.job.count',
     NAMESPACE + '.job.failed',
     NAMESPACE + '.job.succeeded',
     # vpa
@@ -112,8 +108,6 @@
     NAMESPACE + '.vpa.uncapped_target',
     NAMESPACE + '.vpa.upperbound',
     NAMESPACE + '.vpa.update_mode',
-    # namespaces
-    NAMESPACE + '.namespace.count',
 ]
 
 TAGS = {
@@ -148,9 +142,6 @@
         'namespace:kube-system',
     ],
     NAMESPACE + '.pod.count': ['uid:b6fb4273-2dd6-4edb-9a23-7642bb121806', 'created_by_kind:daemonset'],
-    NAMESPACE + '.replicaset.count': ['owner_kind:deployment', 'owner_name:metrics-server-v0.3.6'],
-    NAMESPACE + '.namespace.count': ['phase:active', 'phase:terminating'],
-    NAMESPACE + '.job.count': ['owner_kind:cronjob', 'owner_name:a-cronjob'],
     NAMESPACE
     + '.container.status_report.count.waiting': [
         'reason:containercreating',
@@ -163,14 +154,6 @@
     ],
     NAMESPACE + '.container.status_report.count.terminated': ['pod:pod2'],
     NAMESPACE + '.persistentvolumeclaim.request_storage': ['storageclass:manual'],
-    NAMESPACE
-    + '.service.count': [
-        'namespace:kube-system',
-        'namespace:default',
-        'type:clusterip',
-        'type:nodeport',
-        'type:loadbalancer',
-    ],
     NAMESPACE + '.job.failed': ['job:hello', 'job_name:hello2'],
     NAMESPACE + '.job.succeeded': ['job:hello', 'job_name:hello2'],
     NAMESPACE + '.hpa.condition': ['namespace:default', 'hpa:myhpa', 'condition:true', 'status:abletoscale'],
@@ -373,7 +356,70 @@ def test_update_kube_state_metrics(aggregator, instance, check):
         value=0,
     )
 
-    # deployment counts
+    # services count
+    aggregator.assert_metric(
+        NAMESPACE + '.service.count',
+        tags=['namespace:default', 'type:clusterip', 'optional:tag1'],
+        value=3,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.service.count',
+        tags=['namespace:default', 'type:loadbalancer', 'optional:tag1'],
+        value=2,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.service.count',
+        tags=['namespace:kube-system', 'type:clusterip', 'optional:tag1'],
+        value=4,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.service.count',
+        tags=['namespace:kube-system', 'type:nodeport', 'optional:tag1'],
+        value=1,
+    )
+
+    # namespaces count
+    aggregator.assert_metric(
+        NAMESPACE + '.namespace.count',
+        tags=['phase:active', 'optional:tag1'],
+        value=4,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.namespace.count',
+        tags=['phase:terminating', 'optional:tag1'],
+        value=0,
+    )
+
+    # replicasets count
+    aggregator.assert_metric(
+        NAMESPACE + '.replicaset.count',
+        tags=['namespace:kube-system', 'owner_kind:deployment', 'owner_name:l7-default-backend', 'optional:tag1'],
+        value=1,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.replicaset.count',
+        tags=['namespace:kube-system', 'owner_kind:deployment', 'owner_name:metrics-server-v0.3.6', 'optional:tag1'],
+        value=1,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.replicaset.count',
+        tags=['namespace:kube-system', 'owner_kind:deployment', 'owner_name:kube-dns-autoscaler', 'optional:tag1'],
+        value=1,
+    )
+
+    # jobs count
+    aggregator.assert_metric(
+        NAMESPACE + '.job.count',
+        tags=['namespace:default', 'owner_kind:cronjob', 'owner_name:a-cronjob', 'optional:tag1'],
+        value=1,
+    )
+    aggregator.assert_metric(
+        NAMESPACE + '.job.count',
+        tags=['namespace:default', 'owner_kind:<none>', 'owner_name:<none>', 'optional:tag1'],
+        value=1,
+    )
+
+    # deployments count
     aggregator.assert_metric(
         NAMESPACE + '.deployment.count',
         tags=['namespace:default', 'optional:tag1'],