Conrad monitoring

Train

Load testing

Load testing a storefront with k6 — smoke, load, stress, spike and soak tests, realistic journeys and URL mixes, thresholds, and what to watch while the test runs.
k6 documentationk6 test types

From the k6 line of your Monitoring note.

Know

  • k6: a load testing tool where the test is JavaScript: virtual users run the default function in a loop, stages shape the load, thresholds decide pass or fail, check asserts on responses. Runs locally, in CI or in Grafana Cloud.
  • Test types: smoke (a few users, does it work), load (expected peak, held), stress (raise until it breaks), spike (sudden jump, like a newsletter), soak (normal load for hours, finds leaks).
  • Make it realistic: staging sized like production, real journeys (home, category, product, search, cart), many different URLs so the cache behaves like in real traffic, think time between pages.
  • During the test: watch the dashboards, pod count and scaling, memory, event loop lag, Redis, backend APIs; the result that matters is what gives up first.

Exercise

Exercise
  • Install k6, write a test that ramps up to a peak and holds it, with thresholds on p95 and the error rate, against a local Nuxt app with a cached and an uncached route, and compare p95 and the error rate.
  • Run a short soak test while watching memory in the process, and explain what a leak would look like.

Be able to say

Be able to say· How would you make sure the shop survives a big sale?

"Load test it on staging sized like production with k6: realistic journeys and many different product URLs so the cache behaves like real traffic, ramp up to the expected peak and beyond, and a spike test for the moment the newsletter goes out. Thresholds on p95 and error rate make it pass or fail. While it runs I watch the dashboards to see what gives up first: pods not scaling fast enough, the event loop, Redis or a backend API. Then fix that, pre-scale for the event, and run it again."

Your notes

Open in Notes

Nothing yet. Notes filed here show up on this page and the section overview.

Front-end monitoring

Core Web Vitals (LCP, INP, CLS) with their thresholds and fixes, real user monitoring with Grafana Faro versus Lighthouse, sampling and consent, TTFB on a server-rendered site, and tracing a click from the browser to the database.

Conrad Nuxt

Nuxt questions for the Conrad interview, written from your Nuxt note — server rendering and hydration, runtime and app config, state and memory leaks on the server, and modules, plugins and layers for a platform team.