The Aptible team has performed maintenance on all customer accounts, which has allowed us to implement the following platform improvements:

• Next generation of EC2 hardware: We’ve migrated “General Purpose” containers to the next generation of EC2 hardware for overall better performance

• Implemented CPU Shares and deprecated CPU Isolation (formerly CPU limits): Up until now, customers with CPU Isolation: Disabled have been able to take advantage of additional CPU resources beyond their allocated limit due to traditional CPU Limits not being enabled and the occasional scheduling of containers on larger infrastructure hosts, which allowed for additional CPU usage. We are transitioning to a CPU-share model with right-sized host matching to improve system efficiency and predictable performance. To provide a practical example, a given app could be running a single 4GB container on a 16GB infrastructure host by itself due to several contributing factors. This container would be allocated a single full CPU but could exceed that by up to 4x (400% CPU utilization). With this new model, this container will be rescheduled on a host closer in size to the container. When the container overruns its single full CPU allocation and reaches the limits of the host’s CPU capacity, it would be throttled to ensure the stability of the host and, by extension, the service. This necessary change lays the groundwork for some exciting upcoming changes, including autoscaling.

• Maintenance Commands: We have added two new Aptible CLI commands to simplify further the process of ensuring the required App and Database restarts are complete: aptible maintenance:apps and aptible maintenance:dbs. These commands list Apps and Databases that are required to be restarted to complete any outstanding maintenance and allow customers to track which resources will be restarted by Aptible's SRE time at the indicated maintenance window. Please upgrade your Aptible CLI now to version 0.19.7 or newer to use these commands.

• Improved API Performance