A Volatile Intervals Service determines which requested interval buckets are volatile on a query. Volatile intervals
are intervals which may still not be finalized, and thus should not be cached. The normal use case for this is data
which is being ingested via real time ingestion, and thus which will be replaced later by new historical segments.
Retrieve a simplified sublist of the given intervals aligned to the specified time buckets and that are
partially or fully volatile.
Volatile intervals should not be cached (in the same way that partial intervals are not cached) but are
returned to the user (which partial intervals are typically not). Volatile intervals will typically abut or
overlap incomplete data because they are on the forward edge of ingestion.
granularity - The granularity at which to check for volatility
intervals - The intervals that may contain a volatile sublist
factSource - The fact source table whose intervals are being retrieved