Analyze stress test sampling logs-Performance Testing(PTS)-阿里云帮助中心

During a stress test, sampling logs capture request, response, and timing details for each sampled API call. You can analyze these logs during a test or in a stress test report to quickly identify and resolve issues.

Prerequisites

You have successfully initiated a stress test. For more information, see PTS Stress Testing Quick Start and JMeter Stress Testing Quick Start.

Background information

Sampling logs are collected at a 1% sampling rate and are retained for 30 days.

Procedure

PTS scenario

On the In Progress page or the report details page, click the Request sampling log tab.

Specify the filter conditions and click Search.

Parameter	Description
API for Testing	Select the stress test API.
Request/Business Status	You can filter by request status or business status. The business status indicates whether an assertion passed (success) or failed.
RT Range	Filter logs by response time by specifying a minimum and maximum RT.
HTTP Status Code	Filter logs by status code, such as 200 or 302.

In the search results, find the target log and click View Details in the Actions column.
In the Log Details dialog box, view the following information:
- The General tab displays the headers and body of the request and response.
  
  The page also contains a timing waterfall tab and collapsible sections for General and Exception information. The Exception information section provides a link to Other common errors.
- The Timing Waterfall tab breaks down the timing for each phase of a request's lifecycle.
  - DNS: The time taken to perform a DNS lookup.
  - ConnectTime: The time taken for connection setup.
  - Request Sent: The time taken to send the network request.
  - Waiting (TTFB): The time to first byte. This includes the round-trip network latency and the server's processing time.
  - Content Download: The time taken to receive the response data.
  Note
  - Hover over the timing waterfall to view the values for parameters such as DNS, ConnectTime, Request Sent, and Waiting.
  - The timing data focuses on the core phases of the request lifecycle to help you identify performance bottlenecks.

JMeter scenario

On the In Progress page or the report details page, click the Request Sample Log tab.

Specify the filter conditions and click Search.

Parameter	Description
Sampler	You can select only one Sampler.
Response Status	You can select multiple statuses from the list, including request exception statuses and business statuses (success or failure). The business status is available only for stress test scenarios that have assertions configured.
Time Range	Specify the time when the log was generated. The minimum unit is one second.
RT Range	Filter logs by response time by specifying a minimum and maximum RT.
HTTP Status Code	Filter logs by status code, such as 200 or 302. Note Logs generated before August 14, 2020 cannot be filtered by status code.

In the search results, find the target log and click View Details in the Actions column. The log details, which include trace information, help you identify issues. The details include:
- The General tab displays the detailed fields of the sampling log, including samplerData (such as request method, URL, and cookies), requestHeaders, and responseHeaders.
- The Timing Waterfall tab breaks down the timing for each phase of a request's lifecycle.
  - DNS Lookup: The time taken to perform a DNS lookup.
  - Get Connected: The time taken for connection setup.
  - Request Sent: The time taken to send the network request.
  - Waiting (TTFB): The time to first byte. This includes the round-trip network latency and the server's processing time.
  - Content Download: The time taken to receive the response data.
  Note
  The timing data focuses on the core phases of the request lifecycle to help you identify performance bottlenecks.