Add metrics for fuzzy-finder filtering time

[rafeca]

Summary

In order to understand better the real world performance of the fuzzy-finder and allow us to take better decisions, we want to start logging the time it takes to execute a filtering action.

Along with the time taken to do the filtering, we want to log some additional metadata:

• Number of files in the repository: the time will vary a lot depending on the project size so we need to be able to filter results by similar project sizes.
• Scoring system: Our hypothesis is that the new fast scoring system implemented here will have an important impact on the filtering time, so we want to log which is the scoring type that's being used.

Since the standard scoring system is not controlled by the fuzzy finder (but via atom-select-list) we cannot easily log it from here. That's not a big deal since we're planning on deprecating that system.

Schema of the data

• eventType: fuzzy-finder-v1

• metadata

  field	value
  ec	time-to-filter
  el	Scoring type (alternate or fast)
  ev	Number of files in the project

Sample JSON payload

{
  "eventType": "fuzzy-finder-v1",
  "durationInMilliseconds": 539,
  "metadata": {
    "ec": "time-to-filter",
    "el": "fast",
    "ev": 789,
    "cd2": "x64",
    "cd3": "x64",
    "cm1": 180,
    "cm2": 83,
    "sr": "1680x1050",
    "vp": "1680x591",
    "aiid": "dev"
  },
  "date": "2019-03-21T17:05:32.706Z"
}

Verification process

• [x] Check that Atom is sending the expected JSON payload to the GitHub backend systems by enabling the metrics debug mode.
• [x] Check that the the events get stored on our pipelines correctly.

//cc @telliott27

[rafeca]

Something worth to notice is that this is going to create an event for every single keypress on the fuzzy finder. I'm going to add some logic to throttle the number of events so we send maximum one e.g every minute or so.