1 KRaft Performance Characteristics

Apache Kafka & Strimzi · 4:02

Listen on 93

Lyrics

[Verse 1]
ZooKeeper held us back for years so long
Metadata crawling when we needed strong
KRaft arrived to change the game we play
Milliseconds faster, leading the way
No external systems slowing us down
Built-in controller wearing the crown

[Chorus]
KRaft performance flying high
Latency low, throughput to the sky
Millions of partitions we can scale
Controller speed will never fail
Propagation fast, operations smooth
KRaft performance, feel the groove

[Verse 2]
Remember when a thousand partitions felt like max
ZooKeeper sessions creating cracks
Now we're talking millions without breaking sweat
Partition scalability, the best we've met
Linear growth pattern, predictable and clean
Most efficient cluster that you've ever seen

[Chorus]
KRaft performance flying high
Latency low, throughput to the sky
Millions of partitions we can scale
Controller speed will never fail
Propagation fast, operations smooth
KRaft performance, feel the groove

[Bridge]
Metadata operations per second climbing
Ten thousand requests, perfect timing
Thousand broker clusters running fine
Tuning parameters, everything's aligned
Batch processing keeps the pipeline flowing
Controller efficiency always growing

[Verse 3]
Large-scale tuning needs a careful hand
Metadata cache size, understand
Controller threads scaled to match the load
Network buffers optimized for the road
Replication factor balanced just right
Performance metrics shining bright

[Chorus]
KRaft performance flying high
Latency low, throughput to the sky
Millions of partitions we can scale
Controller speed will never fail
Propagation fast, operations smooth
KRaft performance, feel the groove

[Outro]
From ZooKeeper days to KRaft today
Performance improvements here to stay
Benchmark numbers tell the tale
Modern streaming systems never fail

← 2 ZooKeeper Configuration | 3 KRaft Mode (ZooKeeper-less Kafka) →