From Friday-night deploys to an automated cloud platform
Manual, risky releases replaced by CI/CD, infrastructure as code and a preview of every change.
- Client
- Typical client: software company, 5–20 developers
- Duration
- Typically 4–8 weeks
This is an illustrative example of a typical engagement, not a description of a specific client.
The challenge
Releases happen by hand in the evening, done by the one person who "knows the setup". Servers were configured manually and nobody dares to change them. Production bugs are found late and rarely rolled back quickly.
- Knowledge held by one person
- Environments behave differently
- Rare, large and risky releases
- Rolling back takes hours
How we solve it
- 01
Health check
We review the current setup, risks and bottlenecks, and agree on the target picture.
- 02
Infrastructure as code
Environments are described in Terraform and containers, so test and production match.
- 03
CI/CD
GitHub Actions builds, tests and deploys every change, with a preview for every pull request.
- 04
Safe releases
Small, frequent releases with health checks and one-click rollback. Training and handover to the team.
Before
- Manual evening releases
- Hand-configured servers
- One person knows the setup
- Rollback takes hours
After
- Release on every merge
- Environments described in code
- The whole team can deploy
- Rollback in minutes
Typical impact
- More frequent, smaller releases with lower risk
- New developers get productive faster
- Less dependence on individuals
- Evening work for releases typically disappears
Technology
Recognise the challenge?
Tell us what it looks like on your side and you’ll get a concrete proposal for the next step.