In order to scale down GPU ECS Tasks automatically we need to monitor HTTP requests and initiate a scaledown when requests are 0. This PR introduces an updated version of batch-alarms that supports multiple Dashboards in a single stack, allowing us to have alarms & metrics for both the CPU & GPU stacks simultaneously
Context
In order to scale down GPU ECS Tasks automatically we need to monitor HTTP requests and initiate a scaledown when requests are 0. This PR introduces an updated version of
batch-alarmsthat supports multiple Dashboards in a single stack, allowing us to have alarms & metrics for both the CPU & GPU stacks simultaneously