- Industry
- Banking
- Department
- Engineering
- Team
- Platform and Reliability
- Kind
- Investigator
- Approver
- Head of Technology Operations
- Code
- ENG-19
It designs and runs load, stress and soak tests against the bank's test environment for peak events (salary day, Black Friday card volumes, month-end batch, tax deadline payments) and compares results with the performance budgets and the last baseline. It pins each bottleneck on its component and confirms headroom before peak. **The problem it solves:** Why the services named miss their performance budget at peak: which service and step, at what load, compared with the baseline, and which component limits it.
How it works
6 steps, and a person signs off before step 6.
- 01Looks it up
Finds the record the request is about in the monitoring platform: peak volume history and the test results and resource metrics..
- 02Reads the request
Works out what is being asked and what a complete answer needs.
- 03Checks the facts
Checks figures, dates and rules against the source records, not from memory.
- 04Works it through
Weighs competing explanations against the evidence and shows which one holds up.
- 05Asks a person
Hands the work to Head of Technology Operations with its reasoning before anything goes out.
- 06Acts
Once approved, it acts in the code repository: it posts the performance report on the release. and records what was done.
What it works with
Your knowledge, your systems, your people.
Knows
The performance standard: performance budgets per service, peak volumes (salary day, holiday card spend, month-end), headroom factors, test environment rules (synthetic data only, never production load) and the release gate.
Reads from
the monitoring platform: peak volume history and the test results and resource metrics.
Acts in
the code repository: it posts the performance report on the release.
Hands back
An object with scenario, results (a list with service, p95_ms, error_rate_percent, throughput and vs_baseline_percent), bottlenecks (a ranked list with component, evidence, confidence and fix) and verdict (pass or fail).
Asks a person when
A peak test fails before a known peak date, the result recommends releasing with a regression, or a test would touch production.
Settings
Your limits, set by your managers.
| Setting | Default | What it controls |
|---|---|---|
| p95 budget (ms) | 300 | Responses stay within this. |
| Error-rate budget (%) | 0.5 | Above this fails. |
| Regression tolerance (%) | 10 | Slower than baseline by more fails. |
| Headroom factor | 2 | Capacity needed over forecast peak. |
| Load production | Off | When off, never. |
Limits are settings, not code: a manager changes them without a developer.
Rules it always follows
- Test at 2 times the forecast peak.
- Fail on p95 over 300 ms, errors over 0.5 percent or regressions over 10 percent.
- When off is false, never load production or a live payment system.
- Use synthetic accounts and cards only.
- Pin each bottleneck on its component with evidence.
See the Performance Testing Agent agent on your own work.
We’ll run it on a sample of your work and show you what your approver would see.