The problem
At one digital agency, every release was copied to five cloud servers by hand, by a sysadmin. If no sysadmin was free, the release waited.
Emergency fixes were slow, and investigating problems on the live servers took longer than it should have.
The solution
The answer didn't need a large, complicated platform. It had two simple parts:
- A dashboard where developers submitted what they wanted to deploy.
- A Bash script that read those requests from a MySQL database every five minutes and deployed them to all five servers.
Developers could now push code live and test it themselves, without waiting for anyone.
The result
Releases and emergency fixes no longer depended on who was available. Sysadmins got time back for other work. Developers took more ownership, including shipping fixes out of hours when needed.
The team gave the tool a name of its own, and it stayed in constant use.
What to take from this
- Find the real bottleneck. The problem here wasn't the copying itself. It was that only one group of people could do it.
- Start simple. A dashboard and a script solved the problem. Modern CI/CD tools offer more, and are often the better choice today, but the aim is the same: anyone authorised can release safely.
- Keep it safe. Self-service needs limits. Decide who can deploy, keep a record of what went live and when, and make sure you can roll back.
- Plan for out-of-hours fixes. If one person holds the keys, an evening problem can become a next-morning problem.
This article draws on work by Syed Quadri, SMART5's director, while working at another organisation. It wasn't a SMART5 client project, so the organisation isn't named.