Reduce friction to set up Open Telemetry with Google Cloud backends.

As can be seen with tests/test_automatic.sh, this enables auto-instrumentation like:

opentelemetry-instrument  \
    --configurator=gcp \
    python \
    the/path/to/your/script.py

And with manual instrumentation, it enables very simple logic like:

from opentelemetry.configurator.gcp import OpenTelemetryGcpConfigurator

OpenTelemetryGcpConfigurator().configure()

Compare this with the much more complicated (and incomplete/confusing) instructions found here:

• https://cloud.google.com/trace/docs/setup/python-ot
• https://cloud.google.com/python/docs/reference/spanner/latest/opentelemetry-tracing

Note that in addition to reducing setup friction, this provides a means of unifying configuration steps across different environments as well as across future changes in the ingestion landscape. For example, if the preferred means of ingestion were to change, the API interface here could stay the same, while changing the core implementation logic. (For implementation logic that could break clients, that could come with a change in major version, even though the code interface did not change (making it still very easy for clients to upgrade to the new ingestion path).