Database Scaling

Databases can be resized using the aptible db:restart command, or via the Dashboard.

Database scaling is performed with minimal downtime (usually less than 1 minute).

RAM Scaling

Database Containers can be resized at any time.

Warning

When you resize up, scaling may take a while to complete if Enclave needs to provision additional capacity (i.e. additional EC2 instances). Overall, the process can take up to an hour.

However, Database downtime will always be minimized: Enclave waits until new capacity is available before shutting down your database, so downtime will remain on the order of 1 minute, regardless of how long it takes to provisionin new capacity.

Disk Scaling

Database Disks can be resized at most once a day, and can only be resized up (i.e. you cannot shrink your Database Disk).

I/O Performance

Besides the obvious storage aspect, the size of your Database’s Disk has an impact on its I/O performance. Indeed, Enclave uses AWS gp2 (i.e. general purpose) volumes, which provide a baseline IOPS performance of 3 IOPS / GB of disk.

In addition to the baseline performance, these volumes also offer burst IOPS capacity, which lets you exceed the baseline performance for a period of time.

For production databases, we strongly recommend that you verify that your Database’s I/O activity (see Metrics) remains below the volume’s baseline IOPS performance.

Warning

You should not rely on the volume’s burst capacity during normal activity. Doing so is likely to cause your performance to drop once you exhaust the volume’s burst capacity, which is likely to cause your app to go down as a result.