Choosing Docker Images and Job Layouts

Large pipelines get noisy when every job has a different image and every node looks structurally identical. A better approach is to pair image discipline with layout discipline.

Reuse images intentionally

If a group of jobs shares the same runtime, make that visual pattern obvious. Seeing repeated images across one stage helps reviewers understand the platform assumptions at a glance.

Lay out by stage, not by file order

For visual editing, jobs read better when grouped by stage columns and stacked by function. That keeps build, test, and deploy relationships legible even after the YAML grows.

Keep artifact-heavy jobs larger

Jobs that export artifacts carry more context. They should have enough room to show their artifact list without forcing users to open a second panel just to answer basic questions.