Essays and post-mortems on CI/CD reproducibility, offline-first integration testing, and the cost of slow feedback loops. Cross-posted to Substack and Dev.to, and permanently archived here with RSS.
For a week the largest repository in the workspace was red. Twenty-eight runs out of the last thirty had failed. Not one of them had failed on code — and the only clue was that they all failed in six seconds.
Four containers, 250 MB, 7.9% idle CPU, and 20 seconds before the first request was answered. How one dependency-free Node process replaced the whole arrangement — and why the footprint became a build gate rather than a paragraph.
You do not need a build cache service, a remote execution farm, or a container registry to get a reproducible gate. You need one copy of the setup, and both callers to use it.
Lightweight developer tooling is a sustainability decision disguised as a performance decision. The runtime bill and the energy bill turn out to be the same line item.
Sustainability
Performance
Cost
Tooling
Follow the work as it happens.
New releases, benchmarks and essays. The RSS feed carries every one of them, and the archive keeps the old ones readable.