Guides and reports
Everything we have written about load testing and server sizing, plus real runs with their full numbers.
Real run reports
Actual Stresix runs, with the numbers, what ran out and what the report said to do.
- We load tested Stresix with StresixA real Standard run of Stresix against its own site on a 4 vCPU / 8 GB server: 2,500 simulated users, about 2,600 requests per second, p95 707 ms, 0.26% timeouts, and why the fix was more workers, not a bigger server.
- We load tested Umami, the whole stackA real Standard run of Umami (app and Postgres from its own docker compose file) on a 4 vCPU / 8 GB server: healthy up to about 500 users and 290 requests per second, then timeouts, with one Node thread pinned and the database nearly idle.
Guides
How to size a server and get through a launch without guessing.
- How much server does my web app need?How to find the right server size for a web app or API: estimate peak concurrent users, load test on known hardware, read p95 latency, errors, CPU and memory, then size up only if the numbers say so.
- Will my app survive a launch?How to check that your app holds up before a launch on Hacker News, Product Hunt or Reddit: estimate the spike, test the pages people will actually hit, and fix the cheap things first.
Load testing by setup
What a run does for a Docker image or a plain HTTP API, and how it compares to k6 Cloud.
- Load test a Docker containerHow Stresix load tests a Docker image or a repository: it runs your container on a fresh server, waits for it to answer, then ramps simulated users against it.
- Load test an HTTP endpointRamp simulated users against up to 10 requests to your HTTP app, including POST bodies and auth headers, and watch p95 latency, throughput and error rate per second.
- Stresix vs Grafana k6 CloudStresix vs Grafana k6 Cloud: Stresix needs no script, runs your app on its own server, measures its CPU and memory, and tells you which server size you need. A side-by-side comparison.