This hpa takes scaleup/down actions using the requested CPU of the pods that form the deployment. In adition, for scaledown action it uses custom metrics from the pods. It can be used as an idea to write an HPA based on custom metrics from external sources other than kubernetes API.
Config via the following environment varialbles:
MIN_REPLICAS: Minimum replicas in low activity (CPU). MAX_REPLICAS: Maximum replicas in high activity (CPU). UPPER_CPU_THRESHOLD: Float from 0 to 1. Applied to the total requested CPU to calculate the threshold CPU to scale up. LOWER_CPU_THRESHOLD: Float from 0 to 1. Applied to the total requested CPU to calculate the threshold CPU to scale down. LOOP_TIME_S: Time between scale action evaluation. As a minimum it should be the amout of time that takes to a pod to be ready. NO_SCALE_DOWN_PERIOD: Minimum time a pod will be alive. Used to prevent unstable situations.