Kafka Partition Simulator
Observe how partition distribution, consumer groups, and hot partitions affect throughput and lag.
Scenario presets — set the sliders below to a named difficulty level, then Start
Level 1 — uniform traffic
Level 2 — one tenant is 70%
Level 3 — fewer consumers than partitions
Level 4 — consumer can't keep up
Partitions:
4
Consumers:
4
Producer rate (msgs/sec):
100
Hot key ratio (%):
0
Consumer speed (msgs/sec):
30
▶ Start
■ Stop
↺ Reset
Statistics
Total messages produced
0
Total messages consumed
0
Max partition lag
0
Hottest partition load
0%
Idle consumers
0
Simulation ready. Press Start.
What to observe
Hot key ratio = 0%
: Even distribution. Each consumer handles similar load. Lag stays low.
Hot key ratio > 50%
: One partition receives most messages. One consumer falls behind. Others are idle.
This is a hot partition.
Consumers > Partitions
: Extra consumers are idle. Kafka assigns at most one consumer per partition per group.
Consumer speed > Producer rate
: Lag never accumulates. Consumers keep up.
Consumer speed much lower than Producer rate
: Lag grows. Scale consumers or reduce producer rate.