hpa request continues monitoring via foremast-brain

[pzou1974]

foremast-brain need to continuous check the request until it gets disabled

[pzou1974]

How to make changes:

1. Add timestamp for the historical cache. if current time - cached timestamp > 30 minutes (drive by config) then clear the cache
2. Current metrics --- always start from X minutes (Configurable) before the current time.

Please make sure change (2) with strategy with HPA

[pzou1974]

3. requires golang foremast service change.
   Current and Historical metrics : Need to make start_time and end_time substitute by current real-time starttime and endtime. Please make sure it needs to cover both Prometheus and Wavefront. HPA only by now.
   For non-hpa strategy you only need to make end time substitutable.

[pzou1974]

Note. Historical metric endtime can be current time - 3 minutes while cache TTL expired. start time is endtime - 7 days. Please make 7 days by config.