On April 15, Amazon AWS released an update stating that between 12:40 and 1:43 a.m. Pacific Time, some EC2 instances located in a single availability area (apne1-az4) of the Northeast Asia Pacific Region 1 (AP-NORTHEAST-1) experienced connection problems. This problem stems from the simultaneous interruption of main power supply and backup power supply, which affects the normal operation of relevant EC2 instances. During this period, customers may experience higher error rates and latency on instances launched in the affected area, and other AWS APIs that partially rely on these EC2 instances may also be affected. Our engineers automatically intervened within minutes and immediately started investigating and implementing remediation measures. We don't think this problem will happen again. Only a very small number of instances are still deployed on hardware affected by power outages. Efforts will continue to be made to recover all affected instances and volumes, but for instant recovery, it is recommended to replace these unrecovered instances or storage volumes whenever possible. The issue has now been resolved and the service has returned to normal operation.