Is it possible to somehow send alerts (to email / slack) based on events that occur within a Kubernetes cluster?
In particular, it would be useful to get an alert if a pod has restarted unexpectedly or if a pod cannot start. Similarly it would be useful to know if a pod's CPU usage was over a certain threshold and get an alert.
We have Heapster (with InfluxDB / Grafana backend) installed. While this gives useful data, it unfortunately does not provide us with alerting.
You can check and probably replace InfluxDB with http://prometheus.io (I am not sure about your cases but generally it is possible) which has built-in metrics support.