- Check and return error from NewBatchPoints to prevent runtime panic if
   user provides an unparsable precision time unit in config.
- Provide correct sample config precision examples.
- Update etc/telegraf.conf precision comment.

closes #715

closes #688
closes #394

closes #714

closes #682

* It was still using the previous config name

And on the test configuration file

* Use it as a paramater for the closure

* This should help backwards compatibility when adding more features or
  supported Mesos components

* Fixed import for JSONFlattener{} it's now in parsers, broke after
  rebasing.

* Use timeout as parameter in the http request
* A bit of cleanup
* More tests

* Defined global var for holding default metric groups
* Refactor removeGroup() to work with the whitelist
* Refactor TestRemoveGroup()

* Now the user selects what to push instead of what not
* Required to check and improve tests
* Missing checks in the code when MetricsCol is empty

The plugin will iterate over the Servers slice and create a goroutine
for each of them.

The plugin is able to query a Mesos master and push the metrics, a
blacklist can be configured and a timeout, it's still not used.

Added unit test, might be a good idea to have system test using docker.

this includes:
- Add Accumulator to the Start() function of service inputs
- For message consumer plugins, use the Accumulator to constantly add
  metrics and make Gather a dummy function
- rework unit tests to match this new behavior.
- make "flush_buffer_when_full" a config option that defaults to true

closes #666

closes #650

closes #693

closes #695

closes #191

Add pg_stat_bg_writer stats

closes #683