The AZ Outage DecisionKafka · hard · ~14 min11:40 AM: AWS reports degradation in AZ us-east-1c. Your 9-broker cluster runs 3 brokers per AZ with rack-aware replica placement (RF=3, one replica per AZ, min.insync.replicas=2). All three brokers in 1c just went unreachable.Leadership wants a status call in 15 minutes: are we up, are we losing data, what are our risks, and why is the inter-AZ network bill so high anyway? Have answers.Evidence

Metrics

Cluster Design

Checkpoints

Current blast radius: what's the honest status?

What is the REAL risk to brief while 1c is down?

When 1c returns, what happens — and what should you NOT do during the outage?

Leadership also asks about the inter-AZ network bill. Explain where it comes from in this design and one concrete lever — then wrap the whole incident narrative, interview-style.