Delivery and technology
Performance testing and non-functional quality
How well a team plans, runs and learns from performance and other non-functional testing, from requirements to production.
4 areas · 9 capabilities · 21 questions · answered on a scale from “Agree” to “Disagree”
Maturity levels
- Ad hoc Performance testing happens late, if at all, and depends on one or two specialists.
- Repeatable Key tests can be rerun, but environments, data and results vary from run to run.
- Defined Non-functional requirements, tests and reporting follow an agreed approach across releases.
- Integrated Performance checks run inside the delivery pipeline and results shape decisions before release.
- Continuous Testing and production monitoring feed each other, and performance risk is managed all the time.
What it assesses
Each capability is scored from its questions, and each team or process gets its own level per capability. One example question is shown for each.
Requirements and strategy
Performance, capacity and resilience targets are known and agreed.
“We have measurable performance targets agreed with the business.”
3 questions
Testing effort is planned around risk.
“We have an agreed approach that says which types of test we run (load, stress, soak, failover) and when.”
2 questions
Environments, data and tooling
Tests run somewhere that tells you something about production.
“Our performance test environment is close enough to production that results can be trusted.”
2 questions
Realistic data and traffic.
“Test data has realistic volumes and variety.”
3 questions
Tests are maintainable code, run automatically.
“Performance scripts are kept in version control like any other code.”
2 questions
Execution and analysis
Tests run often enough to catch problems early.
“How often do performance tests run against the main system?”
2 questions
Results are understood and acted on.
“When a test shows a problem, we find the bottleneck rather than just reporting that it was slow.”
3 questions
Production and collaboration
Live behaviour informs testing.
“We compare production performance with what our tests predicted.”
2 questions
Performance is a team concern, not one specialist's.
“Developers, testers and operations work together on performance, rather than handing it over the wall.”
2 questions