Notesfoundations🌱 seedling
Maintainability
Also: Maintainable System · Software Maintenance
Maintainability is how easy a system is to keep running, understand and change over its lifetime. Most of software's cost comes after launch, so it rests on three principles: [[Operability]], simplicity (removing [[Accidental Complexity]]) and [[Evolvability]].
Why it exists
Launch is the cheap part. Fixing bugs, keeping it running, investigating failures, adapting to new platforms and requirements, and repaying old shortcuts go on for years, often long after the original authors have left. Every system built today is tomorrow's legacy system.
How it works
- Operability: routine operations are easy and visible; Observability lets operators see inside.
- Simplicity: good abstractions hide implementation detail, shrinking what each engineer must understand.
- Evolvability: changes ship in small, reversible, compatible steps.
- Because most outages come from changes, how changes ship is part of maintainability: Rolling Upgrades, Canary Releases and Blameless Postmortems.
Trade-offs
| Gains | Costs |
|---|---|
| faster, safer changes for years | up-front investment in tooling, tests and abstraction |
| easier onboarding and on-call | abstractions that are wrong are worse than none |
In practice
- Google SRE: roughly 70% of outages are caused by changes to a live system, so progressive rollout, fast detection and safe rollback are treated as core reliability work.
Pitfalls
- Treating maintainability as polish to add later: complexity compounds, and later rarely comes.
- Optimizing for the first version's speed at the cost of every later change.
Connections
- Part of
- Foundations
- Related
- ReliabilityScalability
Recall
These cards are also in your review deck, scheduled with spaced repetition.