Clarify need of single metadata bucket for Eventing

On our Eventing FAQ page, we say:

To provide better restartability semantics when an Eventing node is offline, some metadata needs to be stored: a Couchbase bucket solves this problem. Setting up the metadata bucket is a one-time activity that is done cluster-wide. It is recommended that the metadata bucket not be used for any other data-storage (which is to say, it should not be accessed by any other application).

This might be better explained by clearly stating that all Eventing functions can share a single metadata bucket; or that all Eventing functions within a cluster can share a Metadata bucket, regardless of the number of functions, or their source and destination buckets.