Add option to use configured host as hostname for esxi hosts (#17544)

* Add option to use configured hostname as hostname for esxi

* Add changelog

* Move logic to main function and add external host tag test

Author: sarah-witt

esxi/assets/configuration/spec.yaml

@@ -42,6 +42,14 @@ files:
           value:
             type: boolean
             example: false
+        - name: use_configured_hostname
+          description: |
+            If true, the check will use the configured `host` parameter for ESXi hostnames instead of the Host name.
+            You may need to use this if you install both the vSphere check and ESXi check to avoid duplicate entries
+            for hosts in the web application UI.
+          value:
+            type: boolean
+            example: true
         - name: collect_per_instance_filters
           description: |
             Use this option to collect metrics tagged with instance values.

esxi/changelog.d/17544.added

@@ -0,0 +1 @@
+Add option to use configured host as hostname.

esxi/datadog_checks/esxi/check.py

@@ -44,7 +44,8 @@ def __init__(self, name, init_config, instances):
         self.host = self.instance.get("host")
         self.username = self.instance.get("username")
         self.password = self.instance.get("password")
-        self.use_guest_hostname = self.instance.get("use_guest_hostname", False)
+        self.use_guest_hostname = is_affirmative(self.instance.get("use_guest_hostname", False))
+        self.use_configured_hostname = is_affirmative(self.instance.get("use_configured_hostname", False))
         self.excluded_host_tags = self._validate_excluded_host_tags(self.instance.get("excluded_host_tags", []))
         self.collect_per_instance_filters = self._parse_metric_regex_filters(
             self.instance.get("collect_per_instance_filters", {})
@@ -238,7 +239,7 @@ def get_available_metric_ids_for_entity(self, entity):
         metric_ids = [vim.PerformanceManager.MetricId(counterId=counter, instance="") for counter in counter_ids]
         return counter_keys_and_names, metric_ids
 
-    def collect_metrics_for_entity(self, metric_ids, counter_keys_and_names, entity, entity_name, metric_tags):
+    def collect_metrics_for_entity(self, metric_ids, counter_keys_and_names, entity, hostname, metric_tags):
         resource_type = type(entity)
         resource_name = RESOURCE_TYPE_TO_NAME[resource_type]
         for metric_id in metric_ids:
@@ -297,7 +298,7 @@ def collect_metrics_for_entity(self, metric_ids, counter_keys_and_names, entity,
                     self.log.debug(
                         "Skipping metric %s for %s because no value was returned by the %s",
                         metric_name,
-                        entity_name,
+                        hostname,
                         resource_name,
                     )
                     continue
@@ -308,7 +309,7 @@ def collect_metrics_for_entity(self, metric_ids, counter_keys_and_names, entity,
                         "Skipping metric %s for %s, because the value returned by the %s"
                         " is negative (i.e. the metric is not yet available). values: %s",
                         metric_name,
-                        entity_name,
+                        hostname,
                         resource_name,
                         list(metric_result.value),
                     )
@@ -321,10 +322,10 @@ def collect_metrics_for_entity(self, metric_ids, counter_keys_and_names, entity,
                         "Submit metric: name=`%s`, value=`%s`, hostname=`%s`, tags=`%s`",
                         metric_name,
                         most_recent_val,
-                        entity_name,
+                        hostname,
                         all_tags,
                     )
-                    self.gauge(metric_name, most_recent_val, hostname=entity_name, tags=all_tags)
+                    self.gauge(metric_name, most_recent_val, hostname=hostname, tags=all_tags)
 
     def set_version_metadata(self):
         esxi_version = self.content.about.version
@@ -388,6 +389,8 @@ def check(self, _):
         }
 
         for resource_obj, resource_props in all_resources_with_metrics.items():
+            resource_type = type(resource_obj)
+            resource_type_name = RESOURCE_TYPE_TO_NAME[resource_type]
 
             if not is_resource_collected_by_filters(resource_obj, all_resources_with_metrics, self.resource_filters):
                 self.log.debug(
@@ -397,16 +400,18 @@ def check(self, _):
 
             hostname = resource_props.get("name")
 
-            resource_type = RESOURCE_TYPE_TO_NAME[type(resource_obj)]
-            if resource_type == "vm" and self.use_guest_hostname:
+            if resource_type == VM_RESOURCE and self.use_guest_hostname:
                 hostname = resource_props.get("guest.hostName", hostname)
 
+            if resource_type == HOST_RESOURCE and self.use_configured_hostname:
+                hostname = self.host
+
             self.log.debug("Collect metrics and host tags for hostname: %s, object: %s", hostname, resource_obj)
 
             tags = []
             parent = resource_props.get('parent')
 
-            if resource_type == "vm":
+            if resource_type == VM_RESOURCE:
                 runtime_host = resource_props.get('runtime.host')
                 runtime_host_props = {}
                 if runtime_host:
@@ -430,7 +435,7 @@ def check(self, _):
             if parent is not None:
                 tags.extend(get_tags_recursively(parent, resource_map))
 
-            tags.append('esxi_type:{}'.format(resource_type))
+            tags.append('esxi_type:{}'.format(resource_type_name))
 
             metric_tags = self.tags
             if self.excluded_host_tags:
@@ -444,7 +449,7 @@ def check(self, _):
             else:
                 self.log.debug("No host name found for %s; skipping external tag submission", resource_obj)
 
-            self.count(f"{resource_type}.count", 1, tags=tags, hostname=None)
+            self.count(f"{resource_type_name}.count", 1, tags=tags, hostname=None)
 
             counter_keys_and_names, metric_ids = self.get_available_metric_ids_for_entity(resource_obj)
             self.collect_metrics_for_entity(metric_ids, counter_keys_and_names, resource_obj, hostname, metric_tags)

esxi/datadog_checks/esxi/config_models/defaults.py

@@ -28,5 +28,9 @@ def instance_ssl_verify():
     return True
 
 
+def instance_use_configured_hostname():
+    return True
+
+
 def instance_use_guest_hostname():
     return False

esxi/datadog_checks/esxi/config_models/instance.py

@@ -78,6 +78,7 @@ class InstanceConfig(BaseModel):
     ssl_capath: Optional[str] = None
     ssl_verify: Optional[bool] = None
     tags: Optional[tuple[str, ...]] = None
+    use_configured_hostname: Optional[bool] = None
     use_guest_hostname: Optional[bool] = None
     username: str
 

esxi/datadog_checks/esxi/data/conf.yaml.example

@@ -39,6 +39,13 @@ instances:
     #
     # use_guest_hostname: false
 
+    ## @param use_configured_hostname - boolean - optional - default: true
+    ## If true, the check will use the configured `host` parameter for ESXi hostnames instead of the Host name.
+    ## You may need to use this if you install both the vSphere check and ESXi check to avoid duplicate entries
+    ## for hosts in the web application UI.
+    #
+    # use_configured_hostname: true
+
     ## @param collect_per_instance_filters - mapping - optional
     ## Use this option to collect metrics tagged with instance values.
     ## Some ESXi metrics can be tagged with instance values.

esxi/tests/test_unit.py

@@ -1198,3 +1198,28 @@ def test_vm_metrics_filters(vcsim_instance, dd_run_check, metric_filters, expect
     aggregator.assert_metric('esxi.host.count')
     aggregator.assert_metric('esxi.vm.count')
     aggregator.assert_all_metrics_covered()
+
+
+@pytest.mark.usefixtures("service_instance")
+def test_use_configured_hostname(vcsim_instance, dd_run_check, aggregator, datadog_agent):
+    instance = copy.deepcopy(vcsim_instance)
+    instance['use_configured_hostname'] = True
+    check = EsxiCheck('esxi', {}, [instance])
+    dd_run_check(check)
+
+    base_tags = ["esxi_url:127.0.0.1:8989"]
+    aggregator.assert_metric("esxi.cpu.usage.avg", value=26, tags=base_tags, hostname="127.0.0.1:8989")
+    aggregator.assert_metric("esxi.mem.granted.avg", value=80, tags=base_tags, hostname="127.0.0.1:8989")
+    aggregator.assert_metric("esxi.host.can_connect", 1, count=1, tags=base_tags)
+
+    datadog_agent.assert_external_tags(
+        '127.0.0.1:8989',
+        {
+            'esxi': [
+                'esxi_datacenter:dc2',
+                'esxi_folder:folder_1',
+                'esxi_type:host',
+                'esxi_url:127.0.0.1:8989',
+            ]
+        },
+    )