DDS Metric Alarm Configuration Suggestions

Scenarios

You can set alarm rules on Cloud Eye to customize the monitored objects and notification policies and keep track of the instance status. DDS allows you to set threshold rules for instance metrics. If the value of a metric exceeds the threshold, an alarm is triggered. The system automatically sends an alarm notification to the cloud account contact through SMN, helping you learn about the running status of your DDS instance in a timely manner.

This section describes how to configure DDS metric alarm rules.

Creating an Alarm Rule

Log in to the management console.
Under Management & Governance, click Cloud Eye.
In the navigation pane, choose Alarm Management > Alarm Rules.
On the displayed Alarm Rules page, click Create Alarm Rule.
On the displayed page, set parameters as prompted.
Pay attention to the following parameters:
- Event Source: Select Document Database Service.
- Dimension: DDS supports instance- and node-level monitoring. Different monitoring metrics support different monitoring dimensions. For details, see DDS Metrics.
  Figure 1 Configuring a monitoring dimension
- For details about other parameters, see Creating an Alarm Rule in the Cloud Eye User Guide.

**Table 1** Suggestions on configuring alarm rules of DDS metrics
Metric ID	Metrics Name	Dimension	Threshold in Best Practices	Alarm Severity in Best Practices	Handling Suggestion
mongo007_connections_usage	Percentage of Active Node Connections	Node	Raw data > 80% for three consecutive periods	Major	Check whether the connection pool is properly used. Check whether the timeout and other parameters are properly configured. For details, see Common Parameter Configuration on the Driver Side. Increase the maximum number of connections. For a replica set instance whose maximum number of connections is less than 16,000, you can increase the maximum number of connections by changing the instance class. For a cluster instance, you can increase the maximum number of connections by adding mongos nodes.
mongo031_cpu_usage	CPU Usage	Node	Raw data > 80% for three consecutive periods	Major	Identify why the CPU usage is high. For details, see How Do I Solve the High CPU Usage Issue? If the CPU usage remains high, upgrade the CPU specifications. For details, see Changing an Instance Class. If the CPU usage remains high and workloads cannot be completely stopped, the specification change may fail. In this case, choose Service Tickets > Create Service Ticket to submit a service ticket and ask engineers to limit the number of connections in the background. After the specification change is complete, change back the number of connections to the original value.
mongo035_disk_usage	Storage Space Usage	Node	Raw data > 80% for three consecutive periods	Major	Identify why the storage usage is high. For details, see High Storage Usage. If the storage usage remains high, scale up storage space. For details, see Scaling Up Storage Space.
mongo032_mem_usage	Memory Usage	Node	Raw data > 90% for three consecutive periods	Major	Identify why the memory usage is high. For details, see High Memory Usage. If the memory usage remains high for a long time, upgrade the instance class. For details, see Changing an Instance Class.
mongo039_avg_disk_sec_per_read	Average Time per Disk Read	Node	Raw data ≥ 0.1s for three consecutive periods	Major	Check whether the instance has performance bottlenecks in CPU, memory, or connections. If yes, solve the bottlenecks based on the related suggestions. If workloads cannot be optimized, upgrade the instance class or choose specifications with better disk performance. For details, see Changing an Instance Class and Instance Specifications.
mongo040_avg_disk_sec_per_write	Average Time per Disk Write	Node	Raw data ≥ 0.1s for three consecutive periods	Major	Check whether the instance has performance bottlenecks in CPU, memory, or connections. If yes, solve the bottlenecks based on the related suggestions. If workloads cannot be optimized, upgrade the instance class or choose specifications with better disk performance. For details, see Changing an Instance Class and Instance Specifications. Perform a primary/secondary switchover on the console.