Looking back, there is one big takeaway: blow things up. Failure injection testing
was the single most valuable exercise because it generated real-world outcomes. The explicit checklist established standards and expected outcomes. Failure injection shined a light on the gaps between real-world outcomes and the expected outcomes documented in the checklist. Teams were surprised by these gaps. These surprises drove the teams to investigate. Therefore, doing coarse-grain failure injections drives learning.
- Everyone learns: Failing big services causes faults across all of the teams and services. ...