Auto Scaling allows your application to always have the compute capacity needed and reduces the need to manually monitor server capacity. You can autoscale based on incoming requests (front-end) or number of jobs in the queue and how long jobs have been in the queue (back-end).
The benefits of Auto Scaling:
Responsive server usage based on traffic load
Handle unexpected traffic spikes and avoid downtime
Plan scaling activities if you have predictable load changes
Uncover and replace faulty instances and unhealthy applications automatically