Feature request: Add short life caching of metrics

[ScrumpyJack]

With the most common redundancy setup being to double scrape the same metrics with a pair of Prometheus servers, it would be nice to see the oracledb_exporter cache its metrics for a configurable amount of time (say 30 or 60 seconds) or possibly even a number of HTTP GETs. ( equal to the number of Prometheus scrapers)