How to send alerts based on Kubernetes / Docker events?

Question

Is it possible to somehow send alerts (via email / slack) based on events that occur in the Kubernetes cluster?

In particular, it would be useful to get a warning if pod unexpectedly restarted or if pod could not start. Similarly, it would be useful to find out if the pod processor was using a specific threshold and get a warning.

We have a Heapster (with support for InfluxDB / Grafana). Although this provides useful data, unfortunately, it does not give us alerts.

Answer 1

Both sysdig and Datadog provide this functionality as well.

Answer 2

You can check and, possibly, replace InfluxDB with http://prometheus.io (I’m not sure about your business, but, as a rule, it’s possible), which I built in support of indicators.