Review operating systems distribution

[philknows]

Expected Behavior

Currently the crawler is pulling a count for the operating systems distribution metric which should be re-evaluated in how it gets this number. This should also be explained on the UI as the large number of unknown makes this metric useless.

If this metric cannot be accurately predicted/scraped, we should remove it from node statistics.

Current Behavior

High number of "unknown" OS distributions. Doesn't make this useful.

Possible Solution

Let's re-analyze how this is captured and see if there is a way to crawl for better metrics in this section.

[spacesailor24]

This is related to my comment from #200. The OS is retrieved from the self-reported UserAgent and set to unknown if it's not provided

[sadiq1971]

This is related to my comment from #200. The OS is retrieved from the self-reported UserAgent and set to unknown if it's not provided

Need to research if there is different way to get os information. If not we can remove the unknown as a type and just do percentage on the valid os only