Skip to content

benchmark-results.md

Latest commit

 

History

History
68 lines (48 loc) · 2.07 KB

benchmark-results.md

File metadata and controls

68 lines (48 loc) · 2.07 KB
Raw

Benchmark Results

This file is intentionally set up as a portfolio artifact. Run the load tests after the database has seeded 100k products, then paste screenshots and before/after numbers here.

Test Environment

  • Machine: TODO
  • .NET SDK/runtime: TODO
  • PostgreSQL version: TODO
  • Redis version: TODO
  • Dataset size: 100,000 products
  • API mode: Release Docker container

Query Comparison

Command:

RATE_LIMITING_ENABLED=false docker compose up -d --build
k6 run --summary-export load-tests/results/search-summary.json load-tests/k6-search-comparison.js

The optimized and slow scenarios run in separate phases so the slow query does not contaminate optimized-query latency.

Endpoint p50 p95 p99 RPS Notes
/api/products/search/slow TODO TODO TODO TODO In-memory filtering after loading active rows
/api/products/search/optimized TODO TODO TODO TODO Keyset pagination, projection, no tracking, indexed predicates

TODO: Add k6 screenshot.

Cache Comparison

Command:

RATE_LIMITING_ENABLED=false docker compose up -d --build
k6 run --summary-export load-tests/results/cache-summary.json load-tests/k6-cache-comparison.js

The cached endpoint is warmed before measurement. Cached and non-cached traffic run in separate phases.

Open http://localhost:8080/benchmarks to view the latest exported results in the API-hosted dashboard.

Endpoint p50 p95 p99 RPS Notes
/api/products/hot/non-cached TODO TODO TODO TODO Database hit on every request
/api/products/hot/cached TODO TODO TODO TODO Redis cache plus ASP.NET response caching

TODO: Add Redis hit-rate screenshot.

PostgreSQL Plans

Capture query plans with:

EXPLAIN (ANALYZE, BUFFERS)
SELECT "Id", "Sku", "Name", "Category", "Price", "InventoryCount", "Rating"
FROM products
WHERE "IsActive" = true
  AND "Category" = 'Laptops'
  AND "Price" >= 100
  AND "Id" > 0
ORDER BY "Id"
LIMIT 51;

TODO: Add slow and optimized query plans.