Bounding Growth

Unchecked accumulation — plugins installed and never pruned, build history retained forever — is the most common way a Jenkins controller slowly degrades.

Every installed plugin is a maintenance liability: another thing to patch, another thing that can break compatibility on a controller upgrade. Install what a pipeline actually needs and prune the rest periodically. Unbounded build history (artifacts, logs, workspace retention) is the most common reason a Jenkins instance slowly runs out of disk — Discard old builds policies, set deliberately per job, prevent this from becoming an emergency.

Related

• Performance Tuning — the concrete resource and retention settings this principle translates into.
• Concourse: Setup & Sizing — the same growth-bounding instinct (garbage collection), a different mechanism.