摘要
目前实时消息的服务能力业务种类增加,业务量大幅提升,对各相关系统的要求不断提高。在系统硬件有限的情况下,一旦短时业务量峰值超负荷或者某业务节点异常,容易引发雪崩效应,导致业务流程整体不可用且难以恢复。经过研究分析,我们提出了一系列方法来提升实时消息服务能力和恢复:各系统接口具备消息队列蓄水能力;增加超时消息预判和下游截断功能;对关键系统设置复制库等。这方法在很大程度上提升消息服务能力,减少了异常状况的影响,便于快速维护,快速定位问题,同时避免影响用户感知和对周边系统造成压力。
at present, the service capacity of real-time information is increasing, the business volume is greatly improved, and the requirements of all related systems are constantly improved. In the case of limited system hardware, it is easy to cause the avalanche effect once the peak load of short-term business volume or abnormal business node can cause the whole business process to be unusable and difficult to recover. Through research and analysis, we propose a series of methods to improve the capability and recovery of real-time message service. Increase timeout message anticipation and downstream truncation function;Set up replication libraries for key systems. This method to a large extent promote message service ability, reduce the influence of abnormal conditions, to facilitate rapid maintenance, rapid positioning problem, at the same time to avoid impact the user awareness and pressure on surrounding system.
出处
《数码设计》
2018年第7期15-16,共2页
Peak Data Science
关键词
实时消息
服务能力提升
超时预判
复制数据
real-time information
Service enhancement
Overtime pre-judgment
Copy the data