Run sheet · MQTT Production Operations: Reference · 15 min
| Time | Teacher does | Learners do | Link |
|---|---|---|---|
| 0–5 min | Show the broker-cluster figure | Trace a publication through another node and identify the saved state | Slide deck |
| 5–7 min | Trace connections, routing and saved state through the broker-cluster figure | Explain why reconnecting through the load balancer does not prove state recovery | Chapter |
| 7–10 min | Work through per-node connections, publication and delivery rates, and memory in the sizing example | Calculate per-node load and identify measurements needed with a node unavailable | Chapter |
| 10–15 min | Ask which state must recover after a node failure | Name one measured limit and one recovery check | Chapter |
If the lab fails: Use the building worksheet and Capacity Planning Metrics table to calculate connections, routing rate and memory on paper. Camera off: Ask pairs to write the per-node publication and delivery rates for the 50,000-device scenario.
Next move: If learners use the worksheet node count as a release limit, revisit surviving-node capacity and measured limits.