Things Fall Apart: Navigating Amazon RDS and Aurora for Over a Decade as a Non-DBA

I triggered storage autoscaling during maintenance, crashed my first Amazon RDS instance on Graviton due to a suboptimal configuration, and ran out of IOPS multiple times.

While attempting to autosmart the automated backups, my storage costs skyrocketed, surpassing any other expenses associated with the RDS instance.

Sometimes things fall apart, and sometimes I succeed. I doubled the DML per second by switching to io2 Block Express volumes. I experimented with various methods to optimize the storage of fragmented tables.

Here are a few lessons from managing dozens of production databases for over a decade, without DBA expertise.