...
Latency improvement of workloads run with acks=all
Baseline latency (ms) | Optimized latency (ms) | Improvement | |
---|---|---|---|
High Partitions | |||
p99 E2E | 188 (183, 193) | 184 (177, 191) | 2.1% |
p99 Produce | 155.65 (153.1, 158.2) | 151.8 (148.3, 155.3) | 2.5% |
Low Partitions | |||
p99 E2E | 393 (402, 384) | 374.5 (349, 400) | 4.7% |
p99 Produce | 390.95 (396.9, 385) | 374.35 (348.4, 400.3) | 4.2% |
Latency improvement of workloads run with acks=1
Baseline latency (ms) | Optimized latency (ms) | Improvement | |
---|---|---|---|
High Partitions | |||
p99 E2E | 106.5 (111, 102) | 101 (104, 98) | 5.2% |
p99 Produce | 84.7 (85.8, 83.6) | 83.3 (82.5, 84.1) | 1.7% |
Low Partitions | |||
p99 E2E | 12.5 (13, 12) | 12.5 (11, 14) | 0% |
p99 Produce | 3.25 (3.3, 3.2) | 2.95 (3, 2.9) | 9.2% |
Workload Details
All tests are run on 6 m5.xlarge Apache Kafka brokers running with Kraft as the metadata quorum in 3 m5.xlarge instances. The clients are 6 m5.xlarge instances running the OpenMessagingBenchmark. The test is run for 70 minutes, during which the brokers are restarted one by one with a 10 minute interval between restarts.
...