* graphite: initial seriesByTag() support

* graphite: groupByTags() function

* graphite: aliasByTags() function

* graphite: fix optional params update, issue #9238

* graphite: fix tag-related functions version

* graphite: add 1.1.x version to config

* graphite: fix seriesByTag() series parsing

ref #9307

Unfortunately CloudWatch dimensions are case-sensitive and it uses both `DBClusterIdentifier` and `DbClusterIdentifier` (notice the lower case `b`) depending on the metric. All metrics which also have the `Role` dimension appear to use `DBClusterIdentifier`, whereas metric with the `EngineName` dimension use `DbClusterIdentifier`.

A 'Github Application' is now something else, and will not work with Grafana.

ref #9279

`bus.AddHandler("sql", SetUsingOrg)` is already called on line 24.

Very minor change.

public/app/plugins/panel/graph/specs/graph_specs.ts: relax tests to be "within" instead of "equal", so they won't fail on i686 (#9286)

grafana_mask_icon.svg is located in the img/ folder.

ref #9279

Summary:
Previously, cloudwatch data sources would only fill in a single null value if
there was missing data. This results in behavior described in #9267. This
resolves that issue by filling the entire missing period with null values. The
null values can then be interpreted as normal by the graphing frontend.

Test Plan:
Used on a data source that had missing data for many consecutive periods.
Ensured that the graph remained at 0 across the entire window.

closes #9234

Fixed some minor writing issues.