Merge branch 'master' of https://github.com/data-mill-cloud/data-mill

Author: pilillo

infrastructure/components/monitoring-stack/README.md

@@ -1,6 +1,6 @@
 # Monitoring Stack
 
-This component installs the CoreOS Prometheus operator, along with an optional [Grafana](https://prometheus.io/docs/visualization/grafana/) dashboard to visualize collected metrics.
+This component installs the [CoreOS Prometheus operator](https://github.com/coreos/prometheus-operator), along with an optional [Grafana](https://prometheus.io/docs/visualization/grafana/) dashboard to visualize collected metrics.
 
 ## Prometheus
 
@@ -15,12 +15,17 @@ either table or graph;
 
 ![Prometheus architecture](https://prometheus.io/assets/architecture.png)
 
+You can also watch this introductory video to have a better overview of Prometheus.
+
+<p align="center">
+<a href="http://www.youtube.com/watch?feature=player_embedded&v=PDxcEzu62jk" target="_blank"><img src="http://img.youtube.com/vi/PDxcEzu62jk/0.jpg" alt="prometheus walkthrough" width="560" height="315" border="1" /></a>
+</p>
+
 ### Data Model
 Prometheus stores all data as time series, i.e. timestamped values grouped under a specific metric name.
 In practice, a metric consists of multiple domains or labels, thus prometheus a metric is a multivariate time series. Metrics can be [exposed](https://prometheus.io/docs/instrumenting/exposition_formats/) using a simple text format.
-A metric has format `<metric name>{<label name>=<label value>, ...} <value> <timestamp>`, where value is a float related to the metric and timestamp is an int64 (milliseconds since epoch).
-Labels enrich the time series by providing context. The name is a UTF-8 string, whereas the value can be of numerical type float as well as `NaN`, `+Inf`, `-Inf`. Naming convention is described [here](https://prometheus.io/docs/practices/naming/). As perceivable, 
-Prometheus does not actually care of labels data types and flattens them into an untyped time series.
+A metric has format `<metric name>{<label name>=<label value>, ...} <value> <timestamp>`, where value is a float related to the metric (as well as `NaN`, `+Inf`, `-Inf`) and timestamp is an int64 (milliseconds since epoch).
+Labels enrich the time series by providing context. The label name and value are UTF-8 strings. Naming convention is described [here](https://prometheus.io/docs/practices/naming/). As perceivable, Prometheus does not actually care of labels data types and flattens them into an untyped time series.
 More complex [metrics](https://prometheus.io/docs/concepts/metric_types/), such as counters, gauges, histograms and summaries, can be defined on data. See [this](https://prometheus.io/docs/instrumenting/exposition_formats/#histograms-and-summaries) example.
 
 By default metrics are stored in a local folder for a period of 15 days. On K8s this folder can be maintained on a persistent volume (PV) or a [StatefulSet](https://kubernetes.io/docs/tutorials/stateful-application/basic-stateful-set/) can be otherwise used to