akua bench
Planned
Benchmark policy evaluation and package render latency.
akua bench [path] [flags]
Uses OPA partial evaluation for policy benchmarks; the KCL interpreter's own timing for package render. Intended for high-throughput evaluators (admission webhooks, CI gates at scale).
Flags
| flag | description | ||
|---|---|---|---|
--iterations=<n> | run each benchmark N times (default 1000) | ||
--input=<file> | use this input for the benchmark (default: workspace defaults) | ||
| `--engine=<auto\ | embedded\ | shell>` | engine selection |
JSON output
{
"benchmarks": [
{
"name": "tier/production:deny",
"iterations": 1000,
"total_ms": 47,
"mean_us": 47,
"p99_us": 82,
"rules_evaluated": 47
}
]
}