Performance Testing for Quality Assurance (QA)s
Performance testing measures how a system behaves under load: response time, throughput, error rate and resource use. Load tests apply expected traffic, stress tests push past limits to find the breaking point, and soak tests run for hours to reveal leaks. Tools include JMeter, k6 and Gatling.
Stage: Non-functional Testing · Part of the Quality Assurance (QA) roadmap · Last reviewed 19 November 2026
What to learn in Performance Testing
- Define a workload model from expected users and requests per second
- Write a k6 or JMeter script with ramp-up stages
- Read percentiles such as p95 and p99 instead of only averages
- Tell load, stress, spike and soak tests apart by goal
- Spot a bottleneck from rising latency alongside CPU or database metrics
- Summarise findings with evidence and a clear pass or fail threshold
Performance Testing interview questions
Why look at p95 latency rather than the average?
The average hides slow outliers. A p95 of 2 seconds means five percent of requests took longer than that, which real users feel. Percentiles show the tail of the distribution where problems appear first under load.
What does a soak test find that a load test does not?
A soak test runs a steady load for hours, so it exposes problems that grow over time: memory leaks, exhausted connection pools, log disk filling or gradual slowdown. A short load test ends before those effects show.
Where Performance Testing fits in the Quality Assurance (QA) roadmap
Performance Testing belongs to the Non-functional Testing stage (beginner level), which takes about 3 hours. Beyond happy paths.
Other units in this stage
Back to the full Quality Assurance (QA) roadmap