Insights is an intelligent O&M tool that performs scheduled inspections on applications. Insights identifies the root causes of events detected during the inspections and provides suggestions on how to rectify the issues that triggered the events. Insights has a variety of built-in inspection modules. You can modify module parameters based on your requirements. This topic describes how to manage inspection modules and modify the parameters of different inspection modules.

Procedure

  1. Log on to the ARMS console .
  2. In the left-side navigation pane, choose Insights > Inspection Configuration.
  3. In the top navigation bar of the MNS console, select the region where your cluster is deployed.
On the Inspection configuration page, you can view the built-in inspection modules. Insights provides the following inspection modules:
  • Inspection module for error rate spikes: The system checks whether error rate spikes are abnormal based on historical data by using the built-in exception detection algorithm. If an abnormal spike in error rate is detected, an exception event is generated and the root causes are analyzed.
  • Inspection module for average response time spikes: The system checks whether average response time spikes are abnormal based on historical data by using the built-in exception detection algorithm. If an abnormal spike in average response time is detected, an exception event is generated and the root causes are analyzed.
  • Inspection module for traffic imbalance: The system checks whether traffic imbalance exists in a specific type of service such as HTTP, Dubbo, and HSF based on the data in the last 15 minutes. By default, the queries on a single server in 5 minutes cannot be lower than 1,000. The difference between the highest QPS and the lowest QPS cannot exceed 30%. If traffic imbalance is detected, an exception event is generated and the root causes are analyzed.

Enable or disable modules

By default, all of the inspection modules are enabled. To enable or disable a module, turn on or off the rounded rectangle in the Enable column of the module on the Inspection configuration page.

If you disable an inspection module, Insights will not inspect the types of events defined in the inspection module. The following table describes the event types that are defined in each inspection module:

Inspection module Event type
Inspection module for error rate spikes
  • Spikes in the error rates of top N interfaces
  • Spikes in the error rates of application services
Inspection module for average response time spikes
  • Spikes in the average response time of top N interfaces
  • Spikes in the average response time of application services
Inspection module for traffic imbalance Traffic imbalance

Modify module parameters

Note The inspection module for traffic imbalance cannot be modified.

On the Inspection configuration page, find the module whose parameters you want to modify and click Edit in the Operation column. In the module parameter modification panel, modify the parameters and click OK.

The following section describes the parameters of each inspection module:

Table 1. Inspection module for error rate spikes
Parameter Description
Overall Application Detection
  • Enabled: The system checks whether the overall error rate of the application is abnormal.
  • Disabled: The system does not check the overall error rate of the application.
Top N Span Detection
  • Enabled: The system filters out the corresponding interfaces based on the specified number of top N interfaces and then checks whether the error rate of each interface is abnormal.
  • Disabled: The system does not check the error rate of each interface.
Additional Spans If the top N interfaces cannot meet your inspection requirements, click Add interface to add the interfaces that you want to inspect. This parameter has the highest priority.
Application blacklist Select the applications that do not need to be inspected.
Interface blacklist Select the interfaces that do not need to be inspected. The priority of this parameter is second only to that of the Additional Spans parameter.
Advanced Settings
Minimum detection RT(ms) When the response time (RT) of the application or interface to be inspected is lower than the set threshold, the system does not conduct inspections.
Minimum detected traffic (qps) When the traffic of the application or interface to be inspected is lower than the set threshold, the system does not conduct inspections.
Minimum Detection Error Rate (%) When the error rate of the application or interface to be inspected is lower than the set threshold, the system does not conduct inspections.
Exception Threshold Setting If the system detects a spike in the error rate and meets any of the following conditions, the system generates an exception event and analyzes the root causes.
  • Exception duration (minutes): The duration an abnormal spike persists before it is considered a valid exception.
  • Increase in error rate: The magnitude by which the error rate increases before it is considered a valid exception.
  • Error Rate Greater Than (%): The percentage by which the error rate exceeds before it is considered a valid exception.
  • The number of errors is greater: The upper limit of the number of errors before it is considered a valid exception.
Table 2. Inspection module for average response time spikes
Parameter Description
Overall Application Detection
  • Enabled: The system checks whether the average response time of the application is abnormal.
  • Disabled: The system does not check the average response time of the application.
Top N Span Detection
  • Enabled: The system filters out the corresponding interfaces based on the specified number of top N interfaces and then checks whether the average response time of each interface is abnormal.
  • Disabled: The system does not check the average response time of each interface.
Additional Spans If the top N interfaces cannot meet your inspection requirements, click Add interface to add the interfaces that you want to inspect. This parameter has the highest priority.
Application blacklist Select the applications that do not need to be inspected.
Interface blacklist Select the interfaces that do not need to be inspected. The priority of this parameter is second only to that of the Additional Spans parameter.
Advanced Settings
Minimum detection RT(ms) When the response time (RT) of the application or interface to be inspected is lower than the set threshold, the system does not conduct inspections.
Minimum detected traffic (qps) When the traffic of the application or interface to be inspected is lower than the set threshold, the system does not conduct inspections.
Exception Threshold Setting If the system detects a spike in the RT and meets any of the following conditions, the system generates an exception event and analyzes the root causes.
  • Exception duration (minutes): The duration an abnormal spike persists before it is considered a valid exception.
  • RT increase: The magnitude by which the RT increases before it is considered a valid exception.
  • RT Greater Than: The upper limit of response time before it is considered a valid exception.