We are using two clusters of Aurora Serverless v2 to host a D2C e-commerce business. On December 3rd, I noticed that one of the clusters had been shut down by recovery for around 10 minutes. According to the event message I discovered,
Recovery of the DB instance has started. Recovery time will vary with the amount of data to be recovered. at 01:57(UTC). after that,
Recovery of the DB instance is complete. at 02:06(UTC).
I discovered that metrics from Performance Insight, mysql-error-running.log, and RDS monitoring had disappeared for 15 minutes, which is appropriate given the situation. Nevertheless, hundreds of SQL queries had been processed normally up to 01:59:18(UTC), as indicated by our DataDog APM metrics. Therefore, there hasn't been a crash or other unusual circumstance to trigger the recovery.
I read this post and got to the point. But I want to make sure that the events I described above can be reproduced and, if possible, how often they can be repeated. I also want to have a theoretical understanding of what might have happened. (I know, I could ask AWS Support Engineering Team if I need an analysis, but in theory.)
