Cross-AZ Disaster Recovery Experiment
Cross-AZ deployment on the cloud is a common disaster recovery architecture. When a single availability zone fails due to power, network, and natural disasters, the disaster recovery availability zone can still provide service for the outside world to ensure service continuity. Through cross-AZ disaster recovery experiments, you can:
Verify the effectiveness of the cross-AZ disaster recovery architecture on the cloud;
Learn about business disaster recovery performance, RTO and RPO;
Verify the effectiveness of the alarm policy when an availability zone fault occurs.
Hybrid Cloud Disaster Recovery Experiment
To improve business disaster recovery capabilities, hybrid cloud deployment has become a common choice for enterprises to migrate to the cloud. Reasonable architecture design can effectively reduce the probability and impact of hybrid cloud service faults. Tencent Smart Advisor-Chaotic Fault Generator provides a convenient experimental method to help you conduct fault experiments in hybrid cloud scenarios. Through this experiment, you can:
Verify the autonomy of services under stress scenarios;
Verify the effectiveness of the hybrid cloud disaster recovery architecture;
Learn about the performance of business faults when Tencent Cloud services are unavailable;
Verify the effectiveness of fault alarms;
Conduct routine fault experiments for hybrid cloud deployment.
Service Stress Experiment
In scenes such as e-commerce activities and flash sales, instantaneous high-concurrency access may cause critical services to face extreme stress or even crash, resulting in page freezes, request failures, and other faults, causing business damage. Chaos engineering experiments can help you discover such disaster recovery potential problems in advance and avoid risks:
Verify the service's autonomy under stress scenes;
Verify the performance of the service under stress scenes;
Verify the effectiveness of service alarms under stress scenes.
Network Fault Experiment
In the gaming scene, a good system architecture has a higher tolerance to weak networks, and can make players basically unaware of ordinary network jitter. It can self-adjust when there is a moderate network fault, and will not be completely unavailable when there is a severe network fault. To ensure user experience, an excellent disaster recovery capability is required for services such as weak network protection, preloading, node acceleration, and concurrent scheduling. Through the experiment in this scene, you can:
Check the disaster recovery capability of the backend modules and clients in network fault scenes.
Verify the effectiveness and instantaneity of abnormal alarms when network faults occur.