Read a sample: 1. Why Rehearse Service Recovery?
The first indication is a message from the service desk: several customers cannot complete an order. A dashboard shows errors, but not enough to explain their cause. The administrator who normally owns the affected application is away, and the operations team is still deciding whether the issue is confined to one component or part of a wider disruption.
In a real incident, people would be trying to diagnose and restore the service while the impact continued. In a tabletop exercise, the team pauses at this moment and talks through what it would do. Participants can examine their assumptions, consult their procedures, and disagree about the next step without making changes to production.
That distinction is the foundation of this book. A tabletop exercise is a structured discussion of a plausible situation, guided by a facilitator and conducted against stated objectives. It is not a live recovery test. Participants do not need to fail over a workload, restore data, disable an account, or simulate production traffic. They discuss decisions and actions as the situation develops, then identify what they know, what they assume, and what needs further verification.
The purpose is not to predict the next disruption. It is to give people a chance to practice working through uncertainty together before an incident makes that work more urgent.
Azure services operate within a larger environment of applications, identities, networks, data, suppliers, and business processes. Even when individual components are well understood, the path from a technical symptom to a business decision may not be obvious. One team may see a failed dependency; another may see a customer-facing outage. Both views matter, but they do not automatically come together.
A written procedure can describe what should happen. A tabletop helps reveal how people interpret that procedure when information is incomplete. It also brings together people who may rarely make decisions in the same room. An administrator may know where to find a recovery runbook, while a business representative knows which transaction cannot wait until morning. During ordinary operations, neither person may need the other’s knowledge. During a disruption, the gap can become consequential.
Rehearsal creates space for questions that are easy to postpone:
- Would the team recognize the situation described in the procedure?
- Who would have enough information to make the next decision?
- What would participants need to know before acting?
- Where would they find that information if their usual tools were unavailable?
- Which statements about recovery are documented, and which are expectations?
These are not trick questions. A useful exercise may end with several answers still unresolved. The value lies in making uncertainty visible while there is time to address it deliberately.
For certification candidates, the same discussion can connect technical concepts to operational judgment. Knowing that a recovery capability exists is different from explaining when a team would consider using it, what assumptions support that choice, and how the outcome would be checked. The scenarios in this book provide opportunities to practice that reasoning without treating a discussion as a substitute for hands-on learning.
A facilitator presents a situation in stages. Participants describe how they would interpret the available information, what they would look for, and what they believe should happen next. An observer records decisions, assumptions, questions, and points of disagreement. The group finishes by discussing what it learned and what should happen after the exercise.
The “table” can be a meeting room or a remote session. The important features are shared context, enough structure to keep the discussion purposeful, and a way to record what emerges. Participants should be able to consult relevant materials if the exercise allows it; remembering every procedure from memory is rarely the point.
Consider the opening order-service scenario. The facilitator might first say that customers are reporting failed orders and that the monitoring view is incomplete. After the team discusses its initial response, the facilitator could add that a dependency owner cannot yet confirm whether its service is healthy.
That additional information is often called an inject: a piece of scenario information introduced to move the discussion forward. An inject can clarify a fact, complicate an assumption, or require a decision under changed conditions. It should serve an exercise objective rather than surprise participants for its own sake.
At each stage, the facilitator can ask participants to make their reasoning visible:
> “What do you know at this point, what are you assuming, and what would you want to check before deciding?”
The question is simple, but the answers can expose important differences. One participant may assume that a recovery option is already available. Another may know that it was planned but never verified. A third may point out that a dependent system has a separate owner. The discussion has already produced something useful, even though no technical action has occurred.
A tabletop is most productive when participants speak in concrete terms. “We would follow the runbook” is a starting answer, not necessarily a complete one. The facilitator might ask which runbook applies, whether it is accessible in the scenario, and what decision it supports. If no one can answer immediately, the observer can record a question for follow-up rather than sending the group into an unplanned document search.
A team can leave a tabletop more prepared to reason about a disruption. It cannot leave with proof that the recovery steps would work simply because participants described them…