Fixes #2099 -- Add GCP Stackdriver support to telegraf.

[derrley]

Required for all PRs:

• Signed CLA.
• Associated README.md updated.
• Has appropriate unit tests. -- not yet.

This is my first PR -- looking to get some feedback on the direction of this. We've got it running in our own environment with mostly successful results, but I know others might have different environments. In the mean time, I'm going to be working on unit tests and beefing up the README. The example config should make the features fairly clear.

[danielnelson]

Thanks for the pull request @derrley, I'll try to take a closer look soon, probably during the 1.7 cycle.

[russorat]

@derrley thanks so much for this PR! Do you mind providing a little more information about your use case and how you use this in your environment?

[JanekLehr]

@russorat
I worked with @derrley while he was developing this plugin. We run a TICK stack in Google Kubernetes Engine, and one of the telegraf instances scrapes Stackdriver metrics and sends them to InfluxDB. Currently the only Kapacitor alert that uses these metrics is one to alert when a pod container was OOMKilled, but it also gives us the ability to track latencies of our HTTP Load Balancer in GCP, and other GCP resource metrics.

By doing this we can have all our alerts in one place (kapacitor) and avoid defining, and paying for, alerts in Stackdriver

[glinton]

Thanks for the PR. Add some tests, a readme, and attend to the requested changes and we'll get this merged.

[glinton]

Maybe either skip the metric, as you currently are, or default it to a chosen type.

[glinton]

@derrley Do you still have time available, and are you interested in continuing this?Thanks

[danielnelson]

Latest work is on #4977