How better CI/CD pipelines reduce client risk

Pipelines are about risk, not speed

A good pipeline doesn’t just automate work, it;

• Reduces risk
• Prevents human error
• Creates stable, predictable releases
• Protects production

Starting simple != Staying simple

Things usually start simple with a project pipeline;

• You build locally
• Upload files to server
• Push to production
• Check site manually, to make sure it hasn’t broken

This can work OK when you’re starting out, and traffic to the site is relatively low. But inevitably one or more of these is going to happen

• Traffic will increase / more eyes on the project
• The build will become more complex

Whether it’s one or both, this will be true: Risk will increase

Small changes can make a big difference

Improvement #1 — compile and build assets in pipeline, not locally

Problem

• Local builds can vary between machines
• Files get forgotten or misplaced
• Environments drift

Solution

Pipeline builds production-ready assets automatically

Benefits

• Consistent builds
• Predictable deployments
• Fewer/no surprises

Improvement #2 — Standardise Your Environment

Problem

• Node versions
• PHP Versions
• Dependancies

Solution

Lock versions to project

Benefits

• Easily to reproduce issues
• Fewer/no pipeline failures

Improvement #3 — Add Deployment Checks

Problem

Errors reach production

Solution

Automated checks

Benefits

• Build succeeds
• Required files are present
• Deployment completes without errors
• Confidence in deployment release

Improvement #4 — Automate Repetitive Tasks

One that all the developers will be familiar with — we’re just applying it to the deployment pipeline.

Problem

Manual steps are open to user error

Solution

Automate where you’re able

Benefits

• Fewer human errors
• Faster deployments
• More consistent results