feat(lib): remove group separator from main user group

[imiric]

This change now generates group tags according to @ppcano's proposal in #948:

{"type":"Point","data":{"time":"2019-08-23T16:00:05.35818107+02:00","value":0.166917,"tags":{"group":"mainGroup","method":"GET","name":"http://httpbin.org/get","proto":"HTTP/1.1","status":"200","url":"http://httpbin.org/get"}},"metric":"http_req_receiving"}
{"type":"Point","data":{"time":"2019-08-23T16:00:05.4684003+02:00","value":1,"tags":{"group":"mainGroup::nestedGroup","method":"GET","name":"http://httpbin.org/get","proto":"HTTP/1.1","status":"200","url":"http://httpbin.org/get"}},"metric":"http_reqs"}

This is a small change, but it seems backwards incompatible to me, in the sense that it will break user scripts that depended on the previous tag names. @na-- @mstoykov Thoughts?

Closes #948

[ppcano]

Yes, I think this will break how the cloud service processes the group tag.

Out of the cloud service, I don't think many users are creating filters based on the group tag, but I may be wrong.

[codecov-io]

Codecov Report

Merging #1119 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master    #1119   +/-   ##
=======================================
  Coverage   73.28%   73.28%           
=======================================
  Files         141      141           
  Lines       10287    10287           
=======================================
  Hits         7539     7539           
  Misses       2304     2304           
  Partials      444      444

Impacted Files	Coverage Δ
lib/models.go	94.52% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c202415...cf93bbb. Read the comment docs.

[na--]

Sorry, since this is a minor breaking change, let's put this on the back-burner a bit, until we have some time to evaluate how it affects the UX in the cloud and in the end-of-test summary, as well as things like this issue: #949

[na--]

Closing because of #948 (comment), we decided to improve the issue from the other side to avoid major breaking changes