- Industry
- Insure AI
- Department
- Engineering
- Team
- Platform and Reliability
- Kind
- Investigator
- Approver
- Release Manager
- Code
- ENG-19
It runs load tests on quote and bind, FNOL and the broker portal before peaks such as renewal waves and expected catastrophe surges, and diagnoses bottlenecks. It forecasts headroom against peak claims volumes. **The problem it solves:** Why the system slows or fails under load: which component, at what load, and what 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 system: the test run metrics and traces..
- 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 Release Manager with its reasoning before anything goes out.
- 06Acts
Once approved, it acts in the incident tool: it records the result and opens fix items. and records what was done.
What it works with
Your knowledge, your systems, your people.
Knows
The performance budgets, peak volumes (renewal season, past catastrophe events), test environments and known bottlenecks such as rating engine calls.
Reads from
the monitoring system: the test run metrics and traces.
Acts in
the incident tool: it records the result and opens fix items.
Hands back
An object with passed, bottlenecks (a ranked list), headroom and fixes.
Asks a person when
Headroom is below 2 before a peak, or a production test is proposed.
Settings
Your limits, set by your managers.
| Setting | Default | What it controls |
|---|---|---|
| p95 budget (ms) | 1200 | Under peak load. |
| Error budget (%) | 1 | Under peak. |
| Regression tolerance (%) | 10 | Against last run. |
| Headroom factor | 2 | Capacity over the worst past peak. |
| Load test in production | Off | When off, never. |
Limits are settings, not code: a manager changes them without a developer.
Rules it always follows
- Pass only under 1200 ms and 1 percent errors.
- Flag regressions over 10 percent.
- Size FNOL for 2 times the worst catastrophe peak.
- When off is false, test only outside production and never with real policies.
- Rank bottlenecks by evidence and say what would confirm each.
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.