Operator Metering 0.7.0

Metering 0.7.0 is primarily intended to provide a snapshot of our current development.
This release changes are focused on improving performance and observability, as well as exposing new ways to use reports.

Below is a list of most major changes:

- Updated CRD API group to metering.openshift.io (breaking change).
- Metering CR now expects a `spec` field nested under each component (reporting-operator, presto, and hfs). See the `manifests/metering-config` directory for an example.
- Updated docker images names to use `metering` nomenclature
- Add support for HTTPs for reporting-operator API.
- Allow protecting reporting-operator API using Openshift auth-proxy on Openshift
- Add support for exposing reporting-operator API using loadbalancer/nodeport services or Openshift routes
- Support configuring node selectors on each component that deploy pods.
- reporting-operator now exposes Prometheus metrics on it's operations such as Report duration, number of metrics imported, etc.
- Improved resource deletion handling.
- Reduces unnecessary reconciles when nothing has changed. Reduces logging verbosity and number of transient errors.
- Add support for using finalizers for ensuring any clean up can be done before a resource is deleted from the API. Currently experimental and default behavior disables this.
- Automated builds now build all things inside Docker, without any on-host requirements other than Docker and Make (make is used to invoke the Docker builds)
- Support reportingStart/reportingEnd on ScheduledReports. This allows backfilling ScheduledReports with past data, and allows for ScheduledReports that do not run forever.
- Initial support for arbitrary inputs into ReportGenerationQueries is available. Currently only used to allow overriding reportingStart/reportingEnd with custom values.
- Initial support for ReportGenerationQueries to directly depend on Reports and ScheduledReports (to allow querying and reporting on other existing results).
- Remove all support for Tectonic