… to {type}-{dataset}-{namespace}

Currently we have the indexing strategy defined as `{type}-{namespace}-{dataset}`. In this PR I propose to change this to `{type}-{dataset}-{namespace}`. As all 3 fields are constant keyword fields, the orders does not matter in most cases.

The reason I propose this change is to better align the name of indices with the name of the other assets:

* ingest pipeline: {type}-{dataset}
* index template: {type}-{dataset}
  * Index pattern: {type}-{dataset}-*
* alias name: {type}-{dataset}-{namespace}

This makes it easier to remember the asset names conventions (at least for me).

It makes a difference when specifying security per namespace: To lock down security, previously it is `/(logs|metrics)-prod-$/` and becomes `/(logs|metrics)-[^-]+-prod-$/`. In any case,w e should help / assist the user to get this right.