Benchmarks

Reproducible evidence without overclaiming.

StreamKernel benchmarks are published as local baselines with explicit profiles, caveats, and room for deployment-specific validation.

Kafka Bench NOOP

956K ops/sec Published local producer ceiling

Kafka ALO WireEvent

525K ops/sec 512-byte payload, at-least-once

mTLS + OPA

366K ops/sec TLSv1.3 + fail-closed policy

MongoDB Insert

163K docs/sec 95.5M document baseline

ONNX -> MongoDB Vector

337 eps 336.8 eps final run, zero record loss

Benchmark matrix

Filter and sort the published baseline profiles.

Use these as evidence paths to reproduce, inspect, and compare against your hardware, payloads, sinks, policies, and enrichment complexity.


Kafka	NOOP producer ceiling	956K ops/sec	At-least-once
Kafka	WireEvent ALO	525K ops/sec	512-byte payload
Kafka	WireEvent EOS	507K ops/sec	Exactly-once
Pulsar	Backlog drain / live producer	Profile captured	Public throughput number pending release approval
Security	mTLS + OPA	366K ops/sec	TLSv1.3 + fail-closed OPA
MongoDB	Insert baseline	163K docs/sec	95.5M docs
AI	ONNX -> Kafka	204 eps	P50 ~298ms, P99 ~400ms
AI	ONNX -> MongoDB Vector	336.8 eps	Run 12 final, zero record loss
AI	MLflow -> Delta	38 eps	Champion artifact at startup

Benchmarks are published as reproducible evidence, not universal deployment guarantees. Current local baselines were captured on an Intel i9-8950HK with CPU-only execution and no GPU inference. Production results depend on hardware, payloads, sink behavior, network topology, JVM configuration, and policy or enrichment complexity.

Commercial path

Want a benchmark review for your pipeline shape?

Compare published profiles against your runtime envelope, sink behavior, payloads, policy path, and AI enrichment requirements.

Request Commercial Demo Discuss OEM Licensing