Back to Home

Performance Analysis

Comprehensive benchmarks from the Argus WAF engine. All tests run on consumer-grade hardware to demonstrate real-world performance characteristics and optimization trade-offs.

Hardware: Intel i5-10210U @ 1.6GHzEnvironment: Go 1.25.5, Linux amd64

2.7

µs

Request Analysis

Custom request analyzer pipeline for telemetry extraction (header parsing, body inspection)

150

ns

Circuit Breaker

Per-call overhead from custom resilience state machine implementation

2.8

µs

Middleware Overhead

Added latency by security layer (WAF integration + context building + body handling)

56%

4-Core Efficiency

Parallel efficiency with shared WAF instance. Lock-free transaction model.

Request Analysis Pipeline

Custom analyzer latency across request types. Minimal requests process in 2.5µs, while complex header parsing reaches 9.1µs. Shows impact of request complexity on telemetry extraction.

Request Processing Time (µs)

Latency Distribution

Middleware latency across security modes. Baseline processing takes 3.5µs, while LatencyFirst adds 2.0µs overhead (5.4µs total). Paranoid mode achieves 4.6µs through optimized rule selection.

Latency Overhead (µs)

Linear Scalability

Parallel WAF processing with per-request transaction isolation. Performance scales from 309µs (1 core) to 133µs (8 cores), achieving 2.3x speedup with lock-free architecture. Efficiency decreases at higher core counts due to shared rule engine contention.

Parallel Scaling

Memory Efficiency

Memory allocation profile per operation. Log truncation achieves 0 B/op for payloads under 4KB through in-place string manipulation. WAF middleware allocates ~6.8KB per request for transaction state and rule matching.

Benchmark NameMemory (Bytes)Allocs
Log Truncation00
Circuit Breaker00
Analyzer1,96814
WAF Check6,75126