Why to move to a managed database

Managing a database is a sensitive, technical endeavor with the risk of a catastrophic downside. The failure of a primary instance can mean downtime for your application and lost money for your company. In the worst case, a faulty backup plan can result in permanent data loss.

Disaster recovery plans are difficult to test given the sensitivity involved. It can be intimidating to truly simulate a disaster scenario on your production database. Accordingly, many operators of self-managed databases don't know how their database will handle an instance failure until it actually happens. Operating a rarely used runbook during the middle of a production incident is a stress-inducing time for all developers.

The cloud has changed how you can handle failure in your applications. The AWS Cloud is split into different AWS Regions, which are geographic locations such as us-east-1 (N. Virginia, USA), us-west-2 (Oregon, USA), or ap-northeast-2 (Seoul, South Korea). Each Region is split into multiple Availability Zones, which are completely separate data centers within the Region.

Each Availability Zone is designed to be isolated from the failure of other Availability Zones in the same Region. Many AWS managed databases take advantage of this isolation by running instances of your database in multiple Availability Zones within a Region. This helps minimize downtime for your end users.

AWS has experience managing large numbers of databases for its customers. Because of this scale, AWS has developed the procedures and processes to manage all sorts of issues with databases. You can easily fail over to a backup instance in the event of an instance failure. You also can automate your database backups to comply with regulations and ensure that your data is safe in the event of an emergency.

When using on-premises data centers, capacity planning is a regular fact of life. You plan months in advance, guessing at your growth rate and hoping for the best. Usually, you overestimate to be safe, wasting money on unused infrastructure. Occasionally, you underestimate and leave your customers unhappy with unexpected downtime.

The public cloud was designed to solve these problems. If you're running your own databases in the public cloud, you can vertically scale by increasing your instance size or horizontally scale by adding additional nodes. But these operations are risky and can result in downtime.

When using managed databases on AWS, you can increase and decrease your database as needed. Your database storage is automatically scaled according to your needs, so you will never run out of disk space. Your database instance size or cluster size can be increased as needed quickly in the console. The resizing is managed for you, without the risk of data loss.

Finally, many commercial databases require purchasing annual licenses up front with complex negotiations. With most AWS managed databases, you choose a pay-as-you-go model that is based only on the instance size and storage used. This results in more flexibility and less guessing up front.