1 comment

[ 388 ms ] story [ 572 ms ] thread
1. Cover your code with tests right from the beginning

2. Realistically simulate what’s happening to the system in the production environment

3. Automate as much as you can

4. Make sure you are simulating a sufficiently long time period

5. Isolate your tests

6. Don’t focus much on microbenchmarks

7. Don’t get caught in the hyper-optimization trap