Kafka in Production Simulator

Researched Model

Entities: producers, topic partitions, brokers, controller quorum, replicas, ISR, consumer group, coordinator, and records.

Rules: producers append to partition leaders; followers replicate; committed writes depend on acks and ISR; consumers pull assigned partitions and commit offsets.

Failure dynamics: broker loss shrinks ISR, can block writes below min.insync.replicas, triggers leader election, and can cause consumer group rebalances.

Parameters

Produce rate

Higher rates increase broker load, consumer lag, and latency.

Broker capacity

Consumer capacity

Replication factor

min ISR

Acks

Consumers

Auto leader rebalance

Scenarios

Events

Cluster healthy Routing: leaders only Group stable

Live Metrics

Throughput

0/s

Errors

0/s

p95 Latency

0ms

Consumer Lag

ISR Health

100%

Rebalances

Partition State

Part	Leader	ISR	Lag	Status

Event Log

Model Notes

Educational simplification: one topic with four partitions, up to three brokers, and a compact KRaft-style controller. Real deployments can contain many topics, partitions, racks, quotas, network links, and client libraries.

Production focus: watch how leader-only writes, follower replication, ISR, min ISR, consumer lag, and rebalances interact during overload and failure.

Visual invariant: every moving event follows the visible connector line and direction. When leaders or assignments change, routes and particles are rebuilt together.