Skip to content

BENCHMARK_RESULTS.md

Latest commit

 

History

History
71 lines (46 loc) · 1.41 KB

BENCHMARK_RESULTS.md

File metadata and controls

71 lines (46 loc) · 1.41 KB
Raw

Benchmark Results

Overview

Performance benchmarks for plans and subscriptions list endpoints.

Running Benchmarks

# Quick run
go test ./internal/handlers/... -bench=. -benchmem

# Full suite with scripts
./scripts/run_benchmarks.sh

# Compare with baseline
./scripts/analyze_benchmarks.sh baseline.txt new.txt

Load Test Smoke Profile

The repository includes a k6 smoke profile that exercises the API with real HTTP requests and authorization. It validates end-to-end service behavior under load and enforces:

  • p95 latency < 250ms
  • error rate < 0.1%
make loadtest-smoke

For remote targets, set LOADTEST_TARGET and JWT_SECRET:

LOADTEST_TARGET=https://staging.example.com \
JWT_SECRET=${JWT_SECRET} \
make loadtest-smoke

Benchmark Categories

1. Dataset Size Tests

  • Empty, Small (10), Medium (100), Large (1K), XLarge (10K)

2. JSON Encoding Tests

  • Isolates serialization performance

3. Full HTTP Tests

  • Complete request/response cycle

4. Parallel Tests

  • Concurrent request handling

5. Filtered Tests

  • Query parameter filtering

Expected Performance

See BENCHMARK_GUIDE.md for detailed baselines and thresholds.

CI Integration

Benchmarks run automatically on PRs to detect regressions.

Analysis

Use benchstat for statistical comparison:

go install golang.org/x/perf/cmd/benchstat@latest
benchstat baseline.txt new.txt