Message Queue

JD Cloud Message Queue is a distributed message queue service independently developed by JD Cloud. The product can provide a series of message cloud services of high reliability, high availability and high processing performance such as message publishing subscriptions, message queries and dead letter queue. It is an indispensable core product in the cloud architecture.

Try Now


High Availability
With the cluster deployment and automatic primary-secondary switching technology, the service availability up to 99.95% is promised.
High Reliability
Raft is introduced to realize the high reliability of data, the data are synchronously written and the data in 3 copies are backed up, with the data reliability of up to 99.999999%, and the default messages can be persistently stored for 3 days and the message at any time point consumed by a consumer ID in 3 days can be reset.
High Performance
Ten million-level of messages can be received, sent and notified per second, without concurrency restrictions, and a huge number of messages are accumulated, with the upper limit of capacity, and the writing delay of messages is controlled within 10ms.
The various messages types are supported, including normal message, sequential message and delay message, so as to meet the demands for different workload scenarios.
Smooth Expansion
To meet the requirements for messaging between workloads, the number of Topic can be expanded elastically and the scale of cluster can be elastically scalable, which is completely transparent to users, and the on-demand billing is adopted, the resource utilization is improved and the costs for redundancy are reduced.
Monitoring Alarm
Various types of monitoring are provided for users, including a variety of visual data monitoring of production TPS, number and size of published messages and number of accumulated messages, etc. The multi-terminal monitoring and early warning can assist you in giving an early warning, prompting risks and quickly positioning and solving problems in advance.


Core Features

Publish Subscription

Producers can create Topic to publish multiple types of messages, and consumers can subscribe to Topic for consumption on demand.

Message View

The users can view message status and content through TopicID and MessageID.

Dead Letter Queue

Messages that cannot be successfully delivered due to various reasons are placed on the Dead Letter Queue which can be viewed and delivered for consumption again.

Reset Consumption Point

Reset the message consumption progress within the time of persistent storage (3 days by default).


Ordinary Message

Solve asynchronous decoupling between businesses, peaking and filling of traffic, message communication, etc.

Sequential Message

Sequential consumption (FIFO) shall be performed in the release order of messages.

Delay Message

The delivery is performed at a specified delay time point after the message-sending time (current time), for example, the delivery is performed 5 minutes after the sending time of the specified message.

Multi-protocol Access

Support HTTP Protocol Access

Support Restful API, easy and convenient to access.

Support TCP Protocol Access

Provide SDK access of a more professional, reliable and stable SDK protocol.

Simple and User-friendly

Smooth Extension

Support elastic expansion which can increase or decrease resources as needed according to service conditions, to improve resource utilization rate and reduce the cost of consumption.

Development Facilitation

Provide a web console, API access interface and java SDK, easy to develop. It is also convenient for service support and service migration.

Security Protection

Comprehensive Monitoring

Provide features such as multi-dimensional monitoring for the running status and performance of resources, stability maintenance and so on. It also provides early warning notification to reduce daily maintenance workload.

Virtual Private Cloud

Instances run in Virtual Private Cloud, enhancing security and isolation. Provide access restrictions such as subnets and IAM policies to protect against network attacks and safeguard your service privacy.


System Decoupling

In the complicated multi-business system of an enterprise, many business processes don’t strongly depend on the core process, and can be placed in the JCQ for JCQ consumers to consume by need, without influencing the core primary process. such as e-commerce scenario (namely such systems as order, shopping cart, logistics and points).

Peak Clipping and Valley Filling

In case of a difference between the upstream and downstream system processing capacity, JCQ will be used as the data buffer to control the information traffic. Consumption will proceed when the downstream system is capable of processing, so as to avoid invalid request and no service capacity. such as sec-killing, panic buying and other scenarios.


The producer has no need to care about the downstream consumer who subscribes the message. The many-to-many broadcasting mode is adopted, and consumption messages may be sent and received simultaneously. For example, message bus and configuration center update.

Asynchronous and Reliable Information Transmission

Receivers may response requests sent by the upstream system after a long time and may not be able to receive such requests because of network, outage, overload and other reasons, while the synchronous writing and three-copy backup of message queue can store the requests to guarantee asynchronous and reliable transmission of messages and timeliness and reliability of requests.