Better alertmanager templates handling

[puffitos]

Description

The alertmanager default template (the rancher one) is hardcoded in the values.yaml, which makes it hard to read, interpret and change. If one wants to edit the template, they must edit the values.yaml and write a new template directly in the YAML file.

Proposal

Use the configMaps value to mount our default as a configMap. The default config of the alertmanager must also be changed to accommodate for this change (https://github.com/prometheus-community/helm-charts/blob/21021a2799659dbb892c29ab661ad8cbb7bb2a1b/charts/kube-prometheus-stack/values.yaml#L338).

The ConfigMap can then be created as a Template in our chart, as the grafana config maps are created.