Alternative way to scrape metrics by exporter

[echerniak]

Current exporter implementation scrapes metrics every 10 seconds (as often, as queue manager publishes them). I suggest to scrape metrics as often, as prometheus requests them and not to depend on queue manager's publishing rate. It will make exporter logic more understandable and simple, but still a lot of corner cases should be discussed. For example, how to behave, if prometheus didn't scrape metrics for a very long time.