feat: upload source maps to Datadog during MFE deploy

[adamstankiewicz]

Related PR: https://github.com/edx/edx-internal/pull/10993

https://docs.datadoghq.com/real_user_monitoring/guide/upload-javascript-source-maps/

Without source maps uploaded to Datadog for MFEs, any reported JS errors are largely not actionable as engineers cannot discern which specific area of code caused the exception due to minified production JS code.

Example from Datadog docs:

Without source maps

JS error without any clues to to which code area/path triggered the error:

With source maps

Human-readable unminified source code of the exact line of JS code that triggered the error, including a direct link to source code on GitHub:

[adamstankiewicz]

[inform/context]

# Can't upload source maps without ``DATADOG_API_KEY``, which must be set as an environment variable
# before executing the Datadog CLI. The Datadog documentation suggests using a dedicated Datadog API key:
# https://docs.datadoghq.com/real_user_monitoring/guide/upload-javascript-source-maps/

It looks like DATADOG_API_KEY exists in edx-internal as an ansible secret (e.g., here), but I'm not too sure whether this is accessible by the frontend configs used in frontend_utils.py.

As is, I'm attempting to extract DATADOG_API_KEY from the MFE APP_CONFIG. Assumes DATADOG_API_KEY is defined under the common frontend config (e.g., here, shared across all MFEs.

[adamstankiewicz]

[inform/context] The frontend_deploy.py script now requires supplying the --common-config-file when it's executed to provide the path to the common frontend config (source).

The related edx-internal PR accounts for this and would be merged first.