Skip to content

BENCHMARKS.md

Latest commit

 

History

History
65 lines (49 loc) · 2.19 KB

BENCHMARKS.md

File metadata and controls

65 lines (49 loc) · 2.19 KB

Agent SRE Performance Benchmarks

Benchmarked on: AMD Ryzen 9 5950X, 32GB RAM, Python 3.11

Summary

Operation Throughput Latency (p50) Latency (p99)
SLO Evaluation 85,000 ops/sec 0.012ms 0.035ms
Error Budget Calc 120,000 ops/sec 0.008ms 0.024ms
Burn Rate Alert 95,000 ops/sec 0.010ms 0.029ms
SLI Recording 200,000 ops/sec 0.005ms 0.014ms
Fault Injection 15,000 ops/sec 0.067ms 0.18ms
Chaos Template Init 8,500 ops/sec 0.12ms 0.31ms
Staged Rollout Analysis 12,000 ops/sec 0.083ms 0.22ms
Rollback Decision 45,000 ops/sec 0.022ms 0.061ms
Full SRE Pipeline 7,200 ops/sec 0.14ms 0.38ms

Key Takeaways

  • Sub-millisecond SRE: Full SLO + chaos + delivery pipeline in <0.4ms p99
  • Real-time burn rate: Alert detection in <30μs
  • 200K SLI recordings/sec: Handle high-frequency agent telemetry

Running Benchmarks

# Run all benchmarks
python -m benchmarks.run_all

# Custom iteration count
python -m benchmarks.run_all --iterations 50000

# Run individual benchmark modules
python -m benchmarks.bench_slo
python -m benchmarks.bench_chaos
python -m benchmarks.bench_delivery

Benchmark Details

SLO Engine

Benchmark Description
SLO Evaluation Full SLO.evaluate() call with indicators and budget checks
Error Budget Calc remaining_percent + burn_rate() computation
Burn Rate Alert Alert detection via firing_alerts()
SLI Recording record() across all 7 indicator types (round-robin)

Chaos Engine

Benchmark Description
Fault Injection inject_fault() event creation and recording
Chaos Template Init instantiate() across all 9 built-in templates
Chaos Schedule Eval Blackout window evaluation with progressive config

Progressive Delivery

Benchmark Description
Staged Rollout Analysis Analysis criteria evaluation (4 metrics)
Rollback Decision check_rollback() with 3 conditions
Traffic Split Calc current_weight + progress_percent computation