Skip to content
at_.DEVELOPMENT

Zero-downtime deploy rig

Forty client sites share one release pipeline that tests every update before it goes live, and undoes it automatically if something breaks — no downtime, no one has to catch it first.

DOCKERGH ACTIONS
40
SITES SHARING ONE PIPELINE
Blue-green
OLD VERSION STAYS LIVE UNTIL THE NEW ONE IS PROVEN
Auto
ROLLED BACK ON FAILED TEST
2024
SHIPPED
01 — THE PROBLEM

Forty sites, one pipeline, zero room for a bad deploy to spread

Deploying forty client sites through a shared pipeline means a single bad release can take all forty down at once, unless the pipeline is built to stop that outright.

02 — WHAT SHIPPED

A release process with a rollback that doesn’t wait for a human to notice

Every deploy runs a full set of automated checks against a live copy of the new version before any real traffic reaches it — the old version stays up the whole time. If a check fails, the release is undone automatically, before it ever reaches a visitor.

Full write-up pending — this project’s placeholder copy is carried over from the original design handoff and will be replaced with a real case study.

40
CLIENT SITES
0
MINUTES OF DOWNTIME PER DEPLOY
Auto
ROLLBACK ON FAILURE

Let's Build Something Cool._

Have a project, a fractional need, or just want to say hi? I read every message myself.