Knowing quickly whether something worked.
WHAT A FEEDBACK LOOP IS
The time between doing something and learning the result.
WHY SHORTER IS BETTER
Problems are cheaper to fix when found immediately.
WHERE LOOPS EXIST
Writing code and running tests Committing and the build result Deploying and knowing it worked Releasing and seeing real behaviour
WHAT TO SHORTEN FIRST
The slowest one that matters.
WHAT MAKES LOOPS SLOW
Tests taking too long to run Manual verification Deployments requiring coordination Monitoring nobody reads
WHAT TO DO ABOUT SLOW TESTS
Run fast checks first, and slower ones separately.
WHAT TO AUTOMATE
Anything checked manually every time.
WHAT TO MONITOR
Whether a deployment succeeded Whether errors rose afterwards
WHY THAT SECOND ONE
A successful deployment that breaks behaviour is a failure nobody noticed.
WHAT TO SET UP
Alerting on error rates, so a bad release announces itself.
WHAT TO MEASURE
Time from problem occurring to someone knowing.
