Problem
Many services are tested with a handful of users and launched to thousands. Slow pages, timeouts and outages appear on the busiest day, in front of everyone.
Outcomes
- A clear view of how the service performs at expected and peak load.
- Bottlenecks found and fixed before launch.
- Known behaviour when parts of the service fail.
- Performance budgets checked on every release.
- Capacity plans based on test results.
Scope
In scope
- Performance requirements and load models.
- Load, stress and soak testing.
- Failure and recovery testing.
- Performance budgets in delivery pipelines.
- Results analysis with the engineering team.
Not included
- Testing against systems you do not own or have permission to test.
Approach
Discovery
We set response time and capacity targets with the business. We build load models from expected traffic and user journeys.
Delivery Planning
We plan the test environment, the load, stress and failure tests, and the approvals needed to test safely.
Build and Test
We run load, stress and failure tests in a safe environment, and work with your engineers to find and fix the causes. We retest after each fix. Before go-live, you get the test results, and a named person decides whether to go ahead.
Operational Readiness, Hypercare and Closure
We add performance checks to the pipeline so regressions are caught, and hand over the scripts, results and capacity findings. We provide extra hypercare for the first few weeks, then close the work with your sign-off.
Deliverables
Technologies
Controls
Frequently Asked Questions
When should performance testing happen?
Early checks during the build, and a full round before launch or any major change in traffic.
Can you test our live service?
Only with written approval and a rollback plan. We usually test in an environment that mirrors live.
What is failure testing?
Deliberately making parts of the service fail, in a safe environment, to check it recovers the way it should.
Do you fix the performance problems you find?
We work with your engineers to find the cause and recommend fixes, and we retest once they are made.
Worried about your busiest day?
A Technical Audit reviews your performance and reliability risks in 10 to 15 working days and gives you a prioritised remediation plan.