Load Phases and Performance Thresholds

Load Phase Types

A performance test consists of phases that simulate different traffic patterns:

"phases": [
  {"duration": 60,  "arrivalRate": 5,  "name": "Warm up"},
  {"duration": 120, "arrivalRate": 20, "name": "Sustained load"},
  {"duration": 60,  "arrivalRate": 50, "name": "Peak load"},
  {"duration": 30,  "arrivalRate": 5,  "name": "Cool down"}
]

Phase	Purpose	Typical Config
Warm up	Gradual ramp to let caches fill, connections pool	30-60s, low rate
Sustained	Steady production-like load	60-300s, expected traffic
Peak	Stress test at maximum expected load	30-60s, 2-5x normal
Spike	Sudden burst to test resilience	10-30s, 10x normal
Cool down	Gradual decrease to observe recovery	30s, low rate

Performance Thresholds

Define what "passing" means:

"thresholds": {
  "p95ResponseTime": 500,
  "p99ResponseTime": 1000,
  "errorRate": 1,
  "minThroughput": 10
}

Threshold	Description	Typical Values
p95ResponseTime	95th percentile response time (ms)	200-500ms for APIs
p99ResponseTime	99th percentile response time (ms)	500-2000ms
errorRate	Maximum acceptable error percentage	0.1-1%
minThroughput	Minimum requests per second	Depends on capacity

Quality Gates

Performance acceptance criteria can be defined per application:

PUT /api/performance-criteria/{applicationId}
{
  "p95ThresholdMs": 500,
  "p99ThresholdMs": 1000,
  "maxErrorRatePercent": 1.0,
  "zero5xxRequired": true,
  "minThroughputRps": 10,
  "maxRegressionPercent": 20,
  "requireWarmupPhase": true,
  "requireSustainedPhase": true,
  "requirePeakPhase": true,
  "minSustainedDurationSec": 60
}

Each criterion becomes a checklist item (AUTO or MANUAL) that feeds into the readiness score.

Auth is checked before dispatch. If the environment declares auth (a bearer login endpoint) but no credential resolves, the run refuses synchronously with CREDENTIAL_UNRESOLVED rather than dispatching a run that fails every request. And a run that errors 100% never reports PASS — even with no thresholds set (NO_SLO_DECLARED, passed: false).

Results

Performance test results include:

• Response time percentiles (P50, P95, P99)
• Throughput (requests per second)
• Error rate (percentage of failed requests)
• Per-endpoint breakdown (which APIs are slowest)
• Comparison with previous runs (regression detection)

From API Test to Performance Test

See From API Test to Load Test — convert any API scenario to a load test with one click.