The real-time request trace module in the cluster management system can track and monitor the execution of a request in real time and displays the latest detailed information about the request. When requests in your business take a long period of time to be responded, you can use this module to locate the causes for further fault diagnosis and problem analysis. You can also use this feature to detect security vulnerabilities at the earliest opportunity.
Prerequisites
You are logged on to the cluster management system of your Lindorm instance. For more information, see Log on to the cluster management system.
Procedure
In the left-side navigation pane, choose .
Select query conditions, including idc, Group, and Server.
Configure filter conditions for request tracking. The following table describes the filter conditions that you can configure.
Filter condition
Description
Trace
Specifies whether to enable the real-time request tracking feature. Valid values:
ON: Enable the real-time request tracking feature.
OFF (default): Do not enable the real-time request tracking feature.
You can click Edit to modify the value of this condition.
Process time(ms)
The time consumed to process the request. Unit: milliseconds. Default value: 1000.
Response size(Byte)
The amount of data contained in the response for the request. Unit: bytes. Default value: 0.
Table
The table for which the request that you want to track is generated
Encoded region name
The shard in which the table is stored.
Method
The method of the request. Valid values: SELECT, UPSERT, GET, SCAN, and PUT.
Handler type
The type of the request handler. Valid values:
CUSTOM: a collection of all external interfaces called by users to handle requests.
USER_READ: the external interfaces that are called by users to handle read requests, including SELECT, GET, and SCAN.
USER_WRITE: the external interfaces that are called by users to handle write requests, including UPSERT and PUT.
LINDORM_READ: the internal interfaces that are called by Lindorm to handle read requests.
LINDORM_WRITE: the internal interfaces that are called by Lindorm to handle write requests.
INDEX_WRITE: the internal interfaces that are called by Lindorm to handle index writing requests.
Client address
The address of the client that initiates the request.
NoteIf you do not specify filter conditions, all processed request on the select nodes are queried by default.
Click Start Trace. The system automatically tracks the requests that meet the filter conditions and displays the details of each request. You can select scroll or clear the displayed requests based on your requirements.
(Optional) Click Stop Trace to stop tracking.