Add ability to configure ILM deletion policy based on storage

[hartfordfive]

Description

In some case, clusters might be created for tenants with a static amount of disk that can't necessarily be easily scaled up quickly. For example, in a self-managed datacenter, there may be a question of procurement time for hardware and then configuration additional Elasticsearch nodes. In some cases, it may be advantageous to giving cluster owners (aka tenants) the ability to continue consuming logs even if their volume has increased. This might lead to the log owners observing a shorter retention period, but at least they can continue to view new incoming logs. It then becomes more of a burden on the owners of the logs themselves to request and justify an increase in capacity so that they can once again view logs on the cluster for the desired period of time.

I realize there may be some complexities and drawbacks in implementing this although it would be great to see this kind of feature. I'm also aware there is a related feature request opened in the curator project (see issue elastic/curator#1692) although having two different tools (ILM + Curator) actively managing some aspect of the life cycle of an index would just add complexity.

[elasticsearchmachine]

Pinging @elastic/es-data-management (Team:Data Management)