What it is
A Kafka consumer group shares a topic's partitions among its members. Whenever a member joins, leaves, crashes or stops polling, the coordinator reassigns partitions in a rebalance. With the classic eager protocol every member gives up all of its partitions first, so the whole group stops until the new assignment arrives. This simulator replays the JoinGroup, SyncGroup and heartbeat messages and measures the stall, the duplicates and the lag they cause.
When to use it
Use it to see why rolling restarts and autoscaling cause lag spikes, what cooperative rebalancing and static membership change, how range, round-robin and sticky assignors differ in how many partitions move, and why a slow consumer can keep kicking itself out. Compare strategies on the same scenario, then replay one message by message.
Gotchas & limits
This is a teaching model of the classic consumer group protocol with one topic and no brokers. Timer defaults are Kafka-like but illustrative and were not checked against the current Kafka documentation; the newer KIP-848 protocol is not modelled. Stall and lag numbers show the mechanism, not a prediction for your cluster.