update netdata global stats and enable them by default

[ilyam8]

Summary

I noticed that

• We use the netdata.uptime metric in some of our alerts.
• This metric is not collected because it is a part of global_statistics_main() that is disabled by default.
• It is disabled by default because global_statistics_main() produces too many metrics and most of them are only needed/understood by developers.

This PR:

• Move the majority of global stats to extended. Create STATS_GLOBAL_EXT thread:
  • extended global stats
  • workers stats
  • sqlilte3 stats
• Enable "netdata monitoring" by default.

This PR does not:

• Some important metrics (such as the number of dbengine metrics/memory usage) should be enabled by default. They are not currently enabled (they are part of the extended metrics). There are more metrics we need to enable by default (e.g. streaming). cc @stelfrag

Test Plan

Install and check "Netdata Monitoring" section.

Additional Information

For users: How does this change affect me?

[thiagoftsm]

PR is working as expected:

24817 25060 ?        00:00:00 P[idlejitter]
24817 25061 ?        00:00:00 HEALTH
24817 25062 ?        00:00:00 STATS_GLOBAL
24817 25063 ?        00:00:00 STATS_GLOBAL_EX
24817 25064 ?        00:00:00 PLUGINSD
24817 25065 ?        00:00:00 SERVICE
24817 25066 ?        00:00:00 STATSD_FLUSH
24817 25068 ?        00:00:00 WEB[1]

[stelfrag]

Ok we can review more useful stats to enable in a future PR.

Note that if we notice gaps in charts we will consider moving sqlite stats in another thread. Some of the functions that get stats may take longer than expected.

[stelfrag]

LGTM