The card below is the “product under test”. Use guided mode to see the bug list upfront, or blind mode to find them yourself first.
Techniques: Metrics & reporting, Error guessing
System performance report — April 2026
| Endpoint | Avg response | P95 | Error rate |
|---|---|---|---|
| GET /dashboard | 210ms | 380ms | 0.1% |
| POST /checkout | 4,200ms | 12,000ms | 0.8% |
| GET /reports/export | 890ms | 45,000ms | 3.2% |
| GET /search | 95ms | 88ms | 0.0% |
Conclusion: All endpoints are performing within acceptable thresholds. No action required.
Your findings
0 / 0 bugs foundTick each bug as you confirm it on the page above. Progress saves automatically.
Write down every bug you find. Be specific — where it is, what’s wrong, what it should be.
Planted bugs (5)
- A 4.2 second average for checkout is very slow. Industry standard for payment endpoints is under 2 seconds. This is not within acceptable thresholds.
- 12 seconds at P95 means 5% of checkout users wait over 12 seconds. High abandonment risk. Not flagged in the report.
- 45 second P95 and 3.2% errors on the export endpoint is a serious performance and reliability problem — not mentioned in the conclusion.
- The P95 for /search is 88ms but the average is 95ms. P95 must always be ≥ average. This data is contradictory — the metrics may be swapped or corrupted.
- The conclusion dismisses obvious performance issues on checkout and export. Either the thresholds are not defined (so anything is 'acceptable') or the report author didn't read their own data.
Performance data: always check that P95 ≥ average (it must be). Look for high error rates (>1% is a concern). Compare against defined SLAs — if no SLAs exist, that's also a finding.