Skip to content
Vision BridgeTechnologies

Platform Engineering

What growing teams should expect from a CI/CD foundation

A reasonable baseline for a team of five to twenty engineers — and the parts that are usually over-engineered too early.

6 min readPublished

Delivery infrastructure attracts two failure modes: teams that keep deploying by hand long past the point it is safe, and teams that adopt a platform designed for an organisation ten times their size.

A reasonable baseline

  • Every merge to the main branch builds and tests automatically
  • Deployment is a single reviewable action, not a sequence of manual steps
  • Environments are created from code, so staging resembles production
  • Secrets live in a managed store, never in the repository
  • Rollback is documented and has been rehearsed at least once
  • Alerts cover the failures that affect users, and reach a named person

What can wait

Multi-region deployment, service meshes, custom internal developer platforms and elaborate progressive-delivery tooling are rarely the constraint for a team of this size. They add operational surface that someone has to maintain while the actual bottleneck is a flaky test suite or a slow review cycle.

Measure the pipeline, not the tooling

Track how long a change takes from merge to production, how often a deployment fails, and how long recovery takes. Those numbers tell you where to invest far better than a comparison of platforms.

Ownership

Delivery infrastructure that no one owns degrades. Before adding a component, decide who maintains it and who is called when it fails at an inconvenient hour. If there is no answer, choose the simpler option.

Related reading

More field notes


Next step

Recognise this problem?

Describe the workflow. We will assess whether software or automation is the right response before proposing a build.