Numbers you can defend in a room full of developers
Anyone can produce a graph. The job is knowing which number matters, what load was realistic, and why the ninety-fifth percentile tells you something the average hides. This page is the route through the JMeter material, the reports to study and the questions you will be asked about them.
Know what you are measuring before you measure it
- Response time, throughput, error rate – and which one the business cares about
- Average versus the 90th and 95th percentile, and why the average lies
- Load, stress, soak and spike – four different questions, four different runs
- Think time and ramp-up: the difference between a test and a denial of service
- What a bottleneck looks like in the numbers before you go looking in the code
The guide
JMeter from a first thread group to a report you would show a stakeholder: samplers, listeners, assertions, correlation, parameterisation and running headless.
Point it at something real
A tutorial that load-tests a public demo site teaches you the tool and nothing about the work. These are ours, so you may hammer them.
The practice API endpoints
Eight applications’ worth of REST endpoints, free and open to everyone, with tokens you can register yourself. Chain a request, vary the data, then put a thread group behind it.
The web applications
Nine practice applications with stable ids on every control, so a recorded script still works tomorrow. Free, no login.
Write it up, then be asked about it
Reports and templates
A performance test case document and the execution and metrics reports are with the QA documents – worth reading before you write your own, because the structure is most of the job.
Interview rounds
JMeter basic, intermediate and advanced, plus the scenario questions – the ones that start “the response time doubled after the release, what do you do first”.