Like PostgreSQL? Like serverless? AWS would like you to like Aurora

Like PostgreSQL? Like serverless? AWS would like you to like Aurora

Amazon has announced PostgreSQL support for its Aurora Serverless database, 11 months after adding MySQL support.

As Amazon puts it, this is “a new deployment option that automatically starts, scales, and shuts down an Amazon Aurora database, and it offers database capacity without the need to provision, scale, and manage any database servers.”

AWS argues this will make capacity planning easier, as rather than running the risk of under or overprovisioning, users simply pay by the second when the database is in use – for example to handle seasonal spikes, or for testing and development. That’s the joy of serverless. Or the danger, if you’re not watching those costs closely.

It’s worth pointing out that the new service is only compatible with PostgreSQL 10.7 – the most recent version is 11.7, while v12.0 is currently in beta. Similarly, AWS’s MySQL support is for v5.6 while the most recent version is 8.x, which debuted last year.

Principal evangelist for serverless at AWS Danilo Poccia explained the setup for Aurora Serverless in a blogpost, which starts with the users setting maximum and minimum capacity. This is defined in Aurora Capacity Units, which are a combination of processing and memory capacity. Pricing for PostgreSQL doesn’t appear to have hit AWS’ website yet, but is $0.07 per ACU hour for the MySQL service, where “1 ACU has approximately 2 GB of memory with corresponding CPU and networking, similar to what is used in Aurora user-provisioned instances.”

“Your client applications transparently connect to a proxy fleet that routes the workload to a pool of resources that are automatically scaled,” he said. “Scaling is very fast because resources are “warm” and ready to be added to serve your requests.”

The minimum storage is 10GB and, based on the database usage, will automatically grow, up to 64 TB, in 10GB increments with no impact to database performance, he continued.