Criterion baseline: f7d5120
Historical suite: this commit predates the platform quarantine. The response-cache, WAF-detection, and double-encoding benchmarks below no longer exist and must not be compared with the current suite.
This is the first evidence-backed Venom microbenchmark baseline. It was captured by the green Quality Metrics workflow run for commit f7d5120.
Criterion results
| Benchmark | Mean | 95% confidence interval |
|---|---|---|
| Response cache hit, 4 KiB | 213.71 ns | 213.00–214.86 ns |
| WAF header detection | 152.28 ns | 151.27–153.99 ns |
| Double URL encoding | 3.113 µs | 3.106–3.124 µs |
Command:
cargo bench --locked -p venom-scanner --bench scanner_benchmarks -- --noplot
The suite took 58 seconds of runner wall time and recorded 550,932 KiB peak RSS. These whole-process values include compilation and harness overhead; they are not per-operation memory measurements.
Release build signal
| Metric | Result |
|---|---|
| Release compile wall time | 43 s |
venom binary size |
4,551,632 bytes |
| Build peak RSS | 579,340 KiB |
Environment and provenance
- Runner: GitHub-hosted
ubuntu-latest - Compiler:
rustc 1.97.1 (8bab26f4f 2026-07-14) - Artifact:
quality-metrics-f7d5120741ae1c2328e1b82b6cec50664a947528 - Artifact digest:
sha256:1a0896de51a3cfa017b36fc32a84b5a20de41b47bc8fbdd25ea7aaf9d0c914c4 - Machine-readable record:
f7d5120.json
The runner CPU model was not captured, so this baseline is suitable for tracking broad regressions on the same workflow class, not for cross-project or capacity comparisons. Endpoint-scale throughput, latency, CPU and memory remain unmeasured.
