Cloud monitoring for kubernetes resource groups

We have a kubernetes cluster where we create individual deployments for each customer. Each deployment is set up with autoscaling.  The current cloud monitoring with Grafana 4 allows us to monitor metrics such as cpu and memory utilization based on a single pod,container and we can group them for the entire cluster but we really need to be able to monitor the cpu/memory utilization per customer deployment, allowing us to gather metrics for pods in a given deployment, preferably based on labels. It would also be nice to be able to monitor hpa activity so we can be alerted when auto-scaling has occurred to help us determine if we need to increase our maximum replicas for a given deployment