Get elastic cluster operation metrics into prometheus

As Elastic Cloud seems to not provide a way to get time series data for internal cluster metrics (we only can get the current state for things like used storage from it’s API) and the limited possibilities of the Elastic monitoring cluster to get a view at the operation of the production cluster aren't really satisfying, we should consider to setup something like https://github.com/justwatchcom/elasticsearch_exporter to get Elastic cluster metrics into Prometheus.

This also would make alerting easier and more standard compared to creating watches within Elastic and we could have dashboards in Grafana.