The KISS principle states that systems are more reliable and understandable when they are kept simple rather than made complicated. KISS stands for "Keep It Simple, Stupid". While this principle was originally intended for aircraft design, it is also very applicable to technical systems and reliability engineering.
Complexity is the enemy of reliability.
More complex systems tend to have more points of failure and more unpredictable interactions between components. Look at Kubernetes 🥲. It requires team(s) of engineers to maintain as it's highly dynamic and complex. Simple designs are easier to understand, troubleshoot, and maintain.
Some ways the KISS principle promotes reliability:
Fewer components means fewer things that can break. Keep it simple and reduce failure modes.
Easy-to-understand systems have fewer human errors. Complex interfaces lead to mistakes.
Simple code has fewer bugs. Complex software is impossible to fully test.
Straightforward documentation makes repair easier. Complicated systems require expertise and certifications to run.
Simple components are easier to make robust. Complex parts break in unexpected ways.
Easy to predict failures in basic designs. Complex systems fail mysteriously.
This is why monoliths are making a come back. Their simplicity improves reliability and improves dev velocity compared to over-engineered microservices.
Another practical example can be goroutines and channels in Golang. While they are powerful concurrency tools, overusing them can lead to hard-to-diagnose deadlocks and race conditions.
Start simple in your designs and code, then add complexity only when absolutely required.