Table of contents
7 Testing Tools and Testing Strategies
9 Test result data and screenshots
9.1 Jmeter performance indicators
If you need a complete report template, remember to find me at the end of the article.
1 Program Overview
Purpose: Find out potential performance defects of the system
Goal: From the perspective of safety, reliability, and stability, find out performance defects, and find out the system's best ability to withstand the number of concurrent users, and the long-running load under the number of concurrent users. If you want to have 100 concurrent users, how to implement Tuning
Overview: This test plan mainly collects and analyzes data related to the processing of concurrent requests by the database for analysis and optimization
Test time: *year*month**day*point*minute-*point*minute
2 References
Related performance test data
3 Terminology Explanation
Performance Testing
English explanation: Performance testing
Concept explanation: Run performance tests to determine system processing capabilities to determine whether the system needs to be optimized
load test
English explanation: Load testing
Concept Explanation: Test the system when it faces multi-resource operation or is attacked
4 System Introduction
Database server, which supports the storage process of the entire system for data
5 test environment
| hardware environment |
Software Environment |
|||||
| server |
server address |
CUP |
Memory |
hard disk |
Software version |
operating system |
| database server |
||||||
6 test indicators
Test time: *year*month*day-*year*month*day
Test scope: When the database processes the server or client request information (insert, query, update, delete) statement, the performance test of the server's various performance indicators
Jmeter indicators: (Because the performance indicators collected by Apache's performance testing tool Jmeter are relatively small, the following data select representative indicators)
1. Average/ms: The average response time of the server processing things (representing the time from the client request to the server processing information and feedback to the client)
2. Throughput/s: The number of requests processed by the server per second (indicating the number of client requests processed by the server per second (unit: one/second))
3. KB/s: The data traffic received by the server per second (indicates the amount of data requested by the server per second received by the client in KB)
Hardware indicators:
1. %Processor time: CPU usage (average less than 75%, better than 50%)
2. System: Processor Queue Length: The number of threads in the CUP queue (average below 2 per processor)
3. Memory: Pages/sec: Number of memory error pages (average less than 20, less than 15 is better)
4. Physical Disk-%Disk Time: Disk usage (average below 50%)
5. SQL Server: Buffer Manager-Buffer Cache Hit Ratio: (The percentage of pages that are found in the buffer cache and do not need to be read from disk, under normal conditions, the ratio exceeds 90%, ideally close to 99%)
7 Testing Tools and Testing Strategies
- Test tool: Apache-Jmeter2.3.2
- Test strategy: Set the database access volume, that is, the number of concurrent users, according to the actual situation within the company and business distribution
- Test data: Because it involves internal company data, it is inconvenient to leak, please forgive me!
- Data description: The selected data are representative data, including stored procedures and queries, updates, deletions, and insertions
8 Test data collection
Collect the results of multiple rounds of tests for comparison, draw a geometric growth graph, and find out the turning point of pressure
9 Test result data and screenshots
Precondition: When the number of users is 80, an error occurs when accessing the database concurrently, so the optimal number of users is set at 75
9.1 Jmeter performance indicators
Average/ms 
- Data analysis :
This graph represents the average response time for a server to process a request,
The best performance is that as the number of concurrent users increases, the average transaction response time is relatively flat.
It can be clearly seen from this figure that as the number of concurrent users increases, the transaction response also increases.
Throughput/s
- data analysis:
This figure shows the number of requests processed by the server per second
The best performance server handles the number of requests that increase with the increase of users
From this figure, it can be seen intuitively that the number of requests processed by the server does not increase as the number of users increases.
KB/S
Database analysis:
- This picture shows the data traffic received by the server per second
- In the best or ideal state, the data traffic received by the server must increase as the number of users increases
- The above figure uses a broken line view to clearly show that when the number of users increases, the request data flow accepted by the server does not increase
Total Requests vs. Number of Users Graph
Database analysis:
- The figure above clearly shows that when 5-15 users initiate requests, the total number of requests is relatively high and smooth
- When the total number of requests after 25-30 is not proportional to the number of concurrent users
- On the contrary, as the number of concurrent users increases, the total number of requests is decreasing!
9.2 Hardware Index Diagram
The figure below shows the hardware information monitoring diagram of the server with 75 concurrent users initiating requests
- data analysis:
The figure above intuitively shows that the average number of memory error pages is 20, and the peak value is as high as 1300 (blue line)
Normal average data is below 20, better below 15
The figure below shows the hardware information monitoring diagram of the server with 50 concurrent users initiating requests
data analysis:
- The figure above intuitively shows that the average number of memory error pages is 20, and the peak value is as high as 1300 (blue line)
- Normal average data is below 20, better below 15
Remarks: (For more hardware indicators, please go to F:\jmeter report\jmeter under the 192.168.1.*** machine)
10 Test conclusion
Jmeter performance index analysis
- It can be seen most intuitively from the Jmeter performance indicators that the network performance is insufficient
Objectively, it can reflect that there is room for optimization of server processing capacity
- Optimization suggestion: increase network speed (increase broadband megabytes)
- 3.5 The server can withstand concurrent access by 75 users, but this test does not represent the server load capacity
Server hardware information monitoring data analysis
- Combined with Jmeter performance indicators and multiple hardware monitoring graphs, it is concluded that memory is one of the server bottlenecks
- Optimization suggestion: improve memory quality, replace larger memory to improve memory processing capacity
Thanks to everyone who read my article carefully! ! !
I personally sorted out some technical materials I have compiled in my software testing career in the past few years, including: e-books, resume modules, various job templates, interview books, self-study projects, etc. Welcome everyone to click on the business card below to get it for free, don't miss it.
Python automated testing learning exchange group: a full set of automated testing interview resume learning materials to obtain Click the link to join the group chat [python automated testing exchange]: 
http://qm.qq.com/cgi-bin/qm/qr?_wv=1027&k=DhOSZDNS -qzT5QKbFQMsfJ7DsrFfKpOF&authKey=eBt%2BF%2FBK81lVLcsLKaFqnvDAVA8IdNsGC7J0YV73w8V%2FJpdbby66r7vJ1rsPIifg&noverify=0&group_code=198408628
