Add in memory Key/Value cache that tasks can access.

It is a common for a task to want to be able to access non time series data. Providing a generic key value cache that tasks can access would fit many of these use cases.

A few uses cases that have come up:
- Checking quotas/thresholds.
- Silencing certain alerts by flagging them as in maintenance mode or the like.

Initial design seems like a key-value store that can be modified via the API and consumed in TICKscripts would be a good starting point.

A few questions:
- Should a task be able to write to the store or only read?
- If a task can write to a store should each task have its own store or should it be global/namespaced?

Seems like this might end up being similar to the Riemann Index, but maybe that is something else altogether. http://riemann.io/concepts.html I'll need to think on this somemore.

No files, tests, or entry points are named, so begin by locating the API and TICKscript task-access paths. The design questions about read/write permissions and store scope must be resolved before done can mean an API-managed key-value store that tasks can consume.

Written by the indexing model from the issue text.