Why RocketMQ Beats Other Open‑Source Queues: A Practical Selection Guide

Selection criteria for an open‑source message queue

Three stakeholder perspectives are considered when evaluating a message‑queue solution without commercial products.

1. Company perspective

Technical cost : server, development and maintenance expenses. Cloud VMs are typical; cost depends on provider performance, discounts, configuration and quantity.

Labor cost : training, recruitment and onboarding. Technologies with steep learning curves increase onboarding time and risk turnover.

Other considerations : large internet companies may initially prefer buying, but cost pressure eventually forces technical optimisation or staff reductions.

2. Middleware team perspective

Stability : essential for reliable business operation.

Feature support : retry, dead‑letter, message tracing, scheduled messages, transactional messages, master‑slave failover, ACL, etc.

Performance : write latency and throughput.

Management platform : easy topic/consumer administration, preferably with an out‑of‑the‑box UI.

Monitoring & alerting : seamless integration with internal monitoring systems such as Prometheus.

Ops, support & community : minimal third‑party dependencies, active open‑source community and responsive support channels.

3. End‑user perspective

Stability : developers need a service that rarely fails.

Migration difficulty : ease of integrating the queue into existing projects.

New project onboarding : simplicity and speed of initial setup.

Compatibility with existing micro‑service frameworks : especially Spring Cloud and other Java ecosystems.

Why RocketMQ satisfies these requirements

Technical cost

RocketMQ runs on modest cloud VMs. A test cluster with 8 CPU, 16 GB RAM, 500 GB SSD (2 master + 2 slave) achieved 40‑50 k TPS send rate, a peak of 200 k TPS consumption, and stabilised at 80‑90 k TPS. Horizontal scaling incurs no noticeable performance loss.

Labor cost

Implemented in Java, RocketMQ provides a stable API with minimal breaking changes across versions, easing upgrades. Java talent is abundant; many large enterprises (e.g., Ant Financial, WeBank, ByteDance, Meituan) already use it, widening the candidate pool.

Stability

Proven in Alibaba’s Double‑11 peak, financial institutions and other high‑traffic services, RocketMQ delivers financial‑grade reliability with years of zero‑incident operation in some deployments.

Rich feature set

Retry and dead‑letter mechanisms

Ordered and transactional messages

Support for tens of thousands of topics

Message filtering and tracing

Automatic master‑slave failover

Native Prometheus metrics

RocketMQ Console for management

ACL‑based access control

Performance

99.9 % of writes finish within 2 ms, far better than many alternatives that exceed 5 ms. Throughput scales horizontally; tests by Didi show comparable TPS to Kafka.

Management platform

RocketMQ Console is provided out‑of‑the‑box:

https://github.com/apache/rocketmq-externals/tree/master/rocketmq-console

Monitoring & alerting

RocketMQ Exporter offers native Prometheus support:

https://github.com/apache/rocketmq-exporter

Ops & community

No Zookeeper dependency – ready to use out of the box.

Active community on DingTalk, WeChat and QQ groups with quick Q&A.

Continuous growth of committers and contributors.

RocketMQ vs. Kafka

Kafka excels in latency‑tolerant, batch‑oriented workloads with moderate topic counts. RocketMQ is better suited for latency‑sensitive, high‑topic scenarios such as financial transaction messages, order‑status updates and push notifications, where millions of topics and sub‑millisecond latency are required.

Key repository links

RocketMQ source code: https://github.com/apache/rocketmq RocketMQ Exporter source: https://github.com/apache/rocketmq-exporter RocketMQ Spring integration: https://github.com/apache/rocketmq-spring C++ client: https://github.com/apache/rocketmq-client-cpp Python client: https://github.com/apache/rocketmq-client-python Go client: https://github.com/apache/rocketmq-client-go RocketMQ Console source:

https://github.com/apache/rocketmq-externals/tree/master/rocketmq-console

Flink connector:

https://github.com/apache/rocketmq-externals/tree/master/rocketmq-flink