Performance testing
To prepare for implementation, consider environment planning, user personas, and performance requirements. Customization and performance are key factors to consider because bad design and customization without performance requirements can negatively impact performance. Common mistakes to avoid include chatty code, retrieving too much data, intercepting synchronous events, and unintended implementations. When developing a plan, consider the following elements:
Real-world scenarios and personas
Average and peak data volumes
Quantified performance expectations
Data preparation activities
Performance testing environments and tools
Documentation tools
Plan for at least two iterations of performance improvements.
The approach to performance testing should be realistic, isolated, use realistic business data, repeatable, functionally correct, and well-documented.
For a successful performance testing plan, make sure that you include realistic scenarios that reflect business criticality and real-world expectations. Test for average and peak data volumes and clearly define measurable expectations for each scenario. Allow time for data preparation and plan for performance testing environments and tools, including licensing and setup time. Document the test results and plan for at least three iterations of performance improvements. This approach ensures a thorough and effective performance testing process.
Optimization iterations
For each optimization iteration, consider the potential benefits and risks of each change. In some cases, making several low-risk tweaks might be enough to meet performance expectations, while in other situations, more complex changes might be necessary but come with extra risks. Make sure that you assess the impact of each change individually and make one change at a time.