Monitoring Clusters Using Cloud Eye

This section describes metrics reported by GES to Cloud Eye as well as their namespaces, lists, and dimensions. You can use APIs provided by Cloud Eye to query the metric information generated for GES.

Namespace

SYS.GES

Monitoring Metrics

**Table 1** GES metrics
Metric ID	Metric	Description	Value Range	Monitored Object
ges001_vertex_util	Vertex Capacity Usage	Vertex usage in a graph instance. The value is the ratio of used vertices to the total vertices. Unit: %	0–100 Type: float	GES instance
ges002_edge_util	Edge Capacity Usage	Edge usage of a graph instance. The value is the ratio of the used edges to the total edges. Unit: %	0–100 Type: float	GES instance
ges003_average_import_rate	Average Import Rate	Average rate of importing vertices or edges to a graph instance Unit: count/s	0–400000 Type: float	GES instance
ges004_request_count	Request Quantity	Number of requests received by a graph instance Unit: count	≥ 0 Type: integer	GES instance
ges005_average_response_time	Average Response Time	Average response time of requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges006_min_response_time	Minimum Response Time	Minimum response time of requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges007_max_response_time	Maximum Response Time	Maximum response time of requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges008_read_task_pending_queue_size	Length of the Waiting Queue for Read Tasks	Length of the waiting queue for read requests received by a graph instance. This metric is used to view the number of read requests waiting in the queue. Unit: count	≥ 0 Type: integer	GES instance
ges009_read_task_pending_max_time	Maximum Waiting Duration of Read Tasks	Maximum waiting duration of read requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges010_pending_max_time_ read_task_type	Type of the Read Task That Waits the Longest	Type of the read request that waits the longest in a graph instance. You can find the corresponding task name in GES documents.	≥ 1 Type: integer	GES instance
ges011_read_task_running_queue_size	Length of the Running Queue for Read Tasks	Length of the running queue for read requests received by a graph instance. This metric is used to view the number of running read requests. Unit: count	≥ 0 Type: integer	GES instance
ges012_read_task_running_max_time	Maximum Running Duration of Read Tasks	Maximum running duration of read requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges013_running_max_time_ read_task_type	Type of the Read Task That Runs the Longest	Type of the read request that runs the longest in a graph instance. You can find the corresponding task name in GES documentation.	≥ 1 Type: integer	GES instance
ges014_write_task_pending_queue_size	Length of the Waiting Queue for Write Tasks	Length of the waiting queue for write requests received by a graph instance. This metric is used to view the number of write requests waiting in the queue. Unit: count	≥ 0 Type: integer	GES instance
ges015_write_task_pending_max_time	Maximum Waiting Duration of Write Tasks	Maximum waiting duration of write requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges016_pending_max_time_ write_task_type	Type of the Write Task That Waits the Longest	Type of the write request that waits the longest in a graph instance. You can find the corresponding task name in GES documents.	≥ 1 Type: integer	GES instance
ges017_write_task_running_queue_size	Length of the Running Queue for Write Tasks	Length of the running queue for write requests received by a graph instance. This metric is used to view the number of running write requests. Unit: count	≥ 0 Type: integer	GES instance
ges018_write_task_running_max_time	Maximum Running Duration of Write Tasks	Maximum running duration of write requests received by a graph instance Unit: ms	≥ 0 Type: integer	GES instance
ges019 _running_max_time_ write_task_type	Type of the Write Task That Runs the Longest	Type of the write request that runs the longest in a graph instance. You can find the corresponding task name in GES documentation.	≥ 1 Type: integer	GES instance
ges020_computer_resource_usage	Computing Resource Usage	Compute resource usage of each graph instance Unit: %	0–100 Type: float	GES instance
ges021_memory_usage	Memory Usage	Memory usage of each graph instance Unit: %	0–100 Type: float	GES instance
ges022_iops	IOPS	Number of I/O requests processed by each graph instance per second Unit: count/s	≥ 0 Type: integer	GES instance
ges023_bytes_in	Network Input Throughput	Data input to each graph instance per second over the network Unit: byte/s	≥ 0 Type: float	GES instance
ges024_bytes_out	Network Output Throughput	Data sent to the network per second from each graph instance Unit: byte/s	≥ 0 Type: float	GES instance
ges025_disk_usage	Disk Usage	Disk usage of each graph instance Unit: %	0–100 Type: float	GES instance
ges026_disk_total_size	Total Disk Size	Total data disk space of each graph instance Unit: GB	≥ 0 Type: float	GES instance
ges027_disk_used_size	Disk Space Used	Used data disk space of each graph instance Unit: GB	≥ 0 Type: float	GES instance
ges028_disk_read_throughput	Disk Read Throughput	Data volume read from the disk in a graph instance per second Unit: byte/s	≥ 0 Type: float	GES instance
ges029_disk_write_throughput	Disk Write Throughput	Data volume written to the disk in a graph instance per second Unit: byte/s	≥ 0 Type: float	GES instance
ges030_avg_disk_sec_per_read	Average Time per Disk Read	Average time per disk read for a graph instance Unit: second	≥ 0 Type: float	GES instance
ges031_avg_disk_sec_per_write	Average Time per Disk Write	Average time per disk write for a graph instance Unit: second	≥ 0 Type: float	GES instance
ges032_avg_disk_queue_length	Average Disk Queue Length	Average I/O queue length of the disk in a graph instance Unit: count	≥ 0 Type: integer	GES instance

Dimensions

Key	Value
instance_id	GES instance

Mapping Between Task Types and Names

**Table 2** Mapping between task types and names
Type	Name
100	Querying vertices
101	Creating a vertex
102	Deleting a vertex
103	Modifying a vertex property
104	Adding a vertex label
105	Deleting a vertex label
200	Querying edges
201	Creating an edge
202	Deleting an edge
203	Modifying an edge property
300	Querying schema details
301	Adding a label
302	Modifying a label
303	Querying a label
304	Modifying a property
400	Querying graph details
401	Clearing graphs
402	Incrementally importing graph data online
403	Creating a graph
405	Deleting a graph
406	Exporting graphs
407	filtered_khop
408	Querying path details
409	Incrementally importing graph data offline
500	Creating a graph backup
501	Restoring a graph from a backup
601	Creating an index
602	Querying indexes
603	Updating an index
604	Deleting an index
700	Running an algorithm

Viewing Instance Monitoring Information

Log in to the GES management console and choose Graph Management.
In the graph list, locate the row that contains the target graph, choose More, and select View Metric to access the Cloud Eye management console. By default, the graph instance monitoring information is displayed.
You can select a monitoring metric name and time range to check the performance curve.

Creating an Alarm Rule

By setting alarm rules for GES, you can customize monitoring objects and notification policies to promptly understand the operational status of GES and serve as an early warning.

Alarm rule settings for GES include parameters such as alarm rule name, monitoring object, monitoring metrics, alarm threshold, monitoring cycle, and notification sending.

This part describes how to set an alarm rule for GES.

Log in to the GES management console and choose Graph Management from the navigation pane on the left.
Locate the row containing the target instance, choose More in the Operation column, and select View Metric to access the Cloud Eye management console and check the GES monitoring information.
Figure 1 Selecting View Metrics

Ensure that the status of the instance whose monitoring information you want to view is Running. Otherwise, you cannot create an alarm.
In the navigation pane on the left of the Cloud Eye management console, choose Alarm Management > Alarm Rules. On the page displayed, click Create Alarm Rule in the upper right corner or in the middle.

On the Create Alarm Rule page, set parameters as prompted.

Setting alarm parameters

Figure 2 Setting parameters

**Table 3** Alarm parameters
Parameter	Description	Example Value
Alarm Type	Alarm type the alarm rule applies to. The value can be Metric or Event.	Metric
Cloud Product	Name of the cloud service the alarm rule is created for	Graph Engine Service - Graph Instances
Resource Level	This parameter is only available when Alarm Type is set to Metric. The options are Cloud product (recommended) and Specific dimension. Take GES as an example. A user purchases a cloud product (GES VMs) and divides the product into multiple sub-dimensions based on metrics, including disks, mount points, and processes.	Cloud product
Monitoring Scope	Resource scope the alarm rule applies to. Select Specified resources and select one or more monitored objects. For GES, select the ID of the cluster instance you have created. Then, set Instance.	Specific resources
Method	There are three options: Associate template, Use existing template, and Configure manually.	Associate template
Template	This parameter is available only when Use template is selected. Select the template to be used. If no alarm template is available, click Create Custom Template to create one that meets your requirements.	-
Alarm Policy	This parameter is available only when Configure manually is selected for Method. Set the policy that triggers an alarm. For example, trigger an alarm if the CPU usage equals to or is greater than 80% for 3 consecutive periods. For details about GES monitoring metrics, see Monitoring Metrics.	-
Alarm Severity	Alarm severity, which can be Critical, Major, Minor, or Informational.	Major

Configure the alarm notification parameters as prompted.

Figure 3 Setting alarm notification parameters

**Table 4** Alarm notification parameters
Parameter	Description	Example Value
Alarm Notification	Whether to send email, SMS, HTTP, or HTTPS notifications to users when an alarm is triggered You can enable (recommended) or disable this function.	Enable this function
Notification Recipient	The options are Notification policies, Notification group, and Topic subscription.	Topic subscription
Notification Policy	If Notification policies is selected for Notification Recipient, you need to select one or more notification policies. You can specify the notification group, template, window, and other parameters in a notification policy. For how to create a notification policy, see Creating, Modifying, or Deleting a Notification Policy.	-
Notification Object	This parameter is mandatory when Notification Recipient is set to Topic subscription. Name of the topic the alarm notification is to be sent to. If you have enabled Alarm Notification, select a topic. If no desirable topics are available, create one first, whereupon the SMN service is invoked. For details about how to create a topic, see Creating a Topic.	SMN topic
Notification Group	This parameter is mandatory when Notification Recipient is set to Notification group. You can select or create a notification group. After creating a notification group, you need to click Add Notification Recipient in the Operation column of the notification group list to add group members and notification methods.	Notification group name
Notification Template	You can select a system template or create a custom notification template.	System template
Notification Window	Notifications are sent only within the notification window specified in the alarm rule. For example, if Notification Window is set to 00:00–08:00, Cloud Eye sends notifications only within this period.	-
Trigger Condition	Condition for triggering the alarm notification. You can select Generated alarm (when an alarm is generated), Cleared alarm (when an alarm is cleared), or both.	-

Click Create. After the alarm rule is created, if the metric data reaches the specified threshold, Cloud Eye will immediately inform you that an exception has occurred.

Transferring Data to OBS

On Cloud Eye, raw metric data is only stored for two days. However, if you subscribe to OBS, you can synchronize the raw data and extend the storage period.

Parent topic: O&M Monitoring and Alarm Reporting

Previous topic: Historical Queries

Next topic: Package Management