Software architecture and its different forms have been discussed in the previous post. The performance matrices are diverse because each layer has unique tasks and functions. The mandatory performance stats for each level to be tracked during the performance test are listed below.
Client Tier
Each client machine application must run more efficiently to produce results faster. The front end usually takes longer to load. As a result, application performance can be improved by reducing front-loading time by 40%. The following are some examples of performance test metrics at the client level:
- HTML source loading time
- TCP connection time
- Image loading speed
- load time for a javascript file
- Response time for HTTP
- HTTP response status
Web Tier
You can study the performance of a web server. through Web Tier statistics. Details how the web server resources were used when performing performance tests. Depending on the type of web server used in the production environment, different web server statistics apply. Microsoft IIS web server performance statistics are:
- Amount of disk time
- Cache hit duration
- Pending requests
- the number of HTTP connections
- Bytes sent and received per second
- Individual bytes
- Bytes with 404 error (not found)
- CPU usage
- Memory usage
- Disk Usage
Apache Web Server Statistics
- CPU load
- Request per second
- Bytes per second
- Bytes per request
- Busy employees
- Inactive employees
Application tier
When it comes to really complicated business applications, the application layer serves as the foundation of the program. WebLogic, Microsoft.Net, Apache TomCat, Web Sphere, J Boss, and other application servers are examples of application servers. The application server performs most of the application’s activities. Therefore, the number of application servers must be taken into account during performance tests. Application server performance statistics are as follows:
- Relationship time
- Membership period
- Complete topics
- Memory usage
- General memory
- Current transactions
- suspended transactions
- Reverse transactions
- Waiting times
- Fault Servlets
- CPU usage
- Memory usage
- Disk Usage
Database layer
At this level, we need to look for queries with long execution times and potential deadlocks. The following factors affect database server performance statistics:
- CPU usage
- Memory usage
- Disk Usage
- Transfer of questions
- query execution speed
- Longer tail
- Connections
- buffer reserve usage
- seconds per page
- Processing period
