Services / DevOps & Cloud Infrastructure

DevOps & Cloud Infrastructure

Release pipelines, hosting and the checks around them, so shipping a change is routine rather than a risk.

Release pipeline: commit, tests, build, deploy and a live check, with a roll back to the previous release if the check fails.Check fails: back to the previous releaseCommitA reviewed changeTestsRun automaticallyBuildSame steps every timeDeployTo the live hostLive checkPages answer as expectedRelease pipeline: commit, tests, build, deploy and a live check, with a roll back to the previous release if the check fails.Check fails: roll backCommitA reviewed changeTestsRun automaticallyBuildSame steps every timeDeployTo the live hostLive checkPages answer as expected

The problem

Releases that depend on one laptop and a list of manual steps. When something breaks, nobody is sure which change caused it.

How we approach it

Every change takes the same path: automated tests, a build, a deploy, and a check that the live site answers. If that check fails, the previous version is one command away.

Typical work

  • Automated build, test and deploy pipelines on GitHub Actions
  • Hosting on Cloudflare Workers, Vercel and Firebase
  • Checks after every release, with a quick way back
  • Build-time asset pipelines that keep hosting costs low

Where you can see it

This website, the one you’re on now, is released the same way: every change is tested, built, deployed to Cloudflare Workers and checked once live.

Tell us what you’re working on.

Describe the problem in your own words. Working out the right approach is part of the job.

Cookies on this site

We use only what the site needs: one cookie that remembers this choice, and, if you allow it, your light or dark theme saved in your browser. No analytics, no advertising, no tracking.

Cookie details