Help Center/ Well-Architected Framework/ Well-Architected Framework and Practices/ Resilience Pillar/ Overload Control/ RES13 Overload Protection/ RES13-03 Overload Detection and Control
Updated on 2025-05-22 GMT+08:00
View PDF
Share

RES13-03 Overload Detection and Control

When an application system is overloaded, the system may be overwhelmed and cannot provide services effectively. Therefore, overload detection and control are required.

  • Risk level

    High

  • Key strategies

    Overload control (also called traffic control) is a process of ensuring that a system, when overloaded, successfully handles all or the agreed number of requests by employing techniques such as rate limiting, degradation, circuit breaking, and elastic scaling. Typical overload control methods are as follows:

    • Rate limiting: The system proactively discards some service requests when it is overloaded.
    • Degradation: When the system is overloaded, lossy services are provided. Measures such as reducing non-critical services and reducing service quality are taken to reduce the system load.
    • Circuit breaking: In a distributed system, if an application fails to invoke a third-party resource or service due to a third-party resource or service fault (including overload), the application stops invoking remote resources or services to prevent the fault from spreading.
  • Related cloud services and tools

    Huawei Cloud provides some cloud services with built-in traffic control capabilities available for configuration.

    • API Gateway: Request throttling policies can be created to control the number of times that an API can be called by a single user or application within a specified period.
    • Cloud Service Engine: Rate limiting can be configured to specify the maximum number of requests that can be accepted within a specified period.