Learn the vocabulary and phrases used to communicate chaos experiment results clearly: "the system degraded gracefully", "we observed unexpected behaviour", "action items from the game day", and more.
0 / 5 completed
1 / 5
An engineer writes in the experiment report: 'The system degraded gracefully.' What does this communicate?
Graceful degradation is a positive finding — it means the system's resilience mechanisms worked. Common examples: 'The recommendation service was injected with latency; the product page loaded without recommendations rather than timing out — graceful degradation confirmed.' Or: 'When we exhausted database connections, the API returned cached responses for read requests — a graceful degradation path.' Contrast with: 'the system failed completely' or 'error rate spiked to 40%' — those indicate ungraceful failure.
2 / 5
How should you communicate 'unexpected behaviour' in a chaos experiment result?
Communicating unexpected behaviour requires specificity: what was injected, what was affected (including unexpected components), what metrics showed the anomaly, and a preliminary root cause hypothesis. Vague reports ('something weird happened') are not actionable. Good unexpected behaviour reports become the most valuable experiment findings — they reveal hidden dependencies, cascading failure paths, and architectural assumptions that were never explicitly tested.
3 / 5
What makes an 'action item from the game day' well-formed and actionable?
Well-formed action items have: a specific technical change (circuit breaker, not 'resilience improvement'), the exact component to be modified, the owner (a named team, not 'someone'), and a deadline (a sprint or date, not 'later'). Vague action items like 'improve resilience' are never completed because no one knows what done looks like. The best game day action items read like engineering tickets — because they should become engineering tickets.
4 / 5
How do you correctly state that a chaos hypothesis was rejected?
Hypothesis rejection is not a failure — it is a finding. The communication must include: the injected condition, the metric that was violated, the specific threshold exceeded, and the preliminary root cause. Saying 'the experiment failed' confuses the experiment outcome (which succeeded in revealing a gap) with system failure. A rejected hypothesis with a clear finding and action item is a highly valuable outcome.
5 / 5
A report states: 'The system recovered within 47 seconds of removing the fault.' Why is this phrasing important?
Measuring and reporting recovery time is a key chaos engineering output. 'The system recovered within 47 seconds' establishes a baseline: if your RTO (Recovery Time Objective) is 30 seconds, this is a gap. If the next experiment after adding a circuit breaker shows recovery in 12 seconds, you have measured improvement. Always specify what 'recovered' means: 'error rate returned to below 0.5% and latency returned to below 300ms within 47 seconds of removing the fault.'