BUG: ECSRun on Run Screen breaks if run_task_kwargs are not provided

[cicdw]

Bug Description

If you don't provide run_task_kwargs when customizing a Run Config on the Run screen, the flow errors out with

botocore.exceptions.ParamValidationError: Parameter validation failed:
Unknown parameter in input: "null", must be one of: capacityProviderStrategy, cluster, count, enableECSManagedTags, enableExecuteCommand, group, launchType, networkConfiguration, overrides, placementConstraints, placementStrategy, platformVersion, propagateTags, referenceId, startedBy, tags, taskDefinition

This is because when no key / value pairs are provided for run_task_kwargs, it seems the UI submits {"null": null} as the key / value pair:

"run_task_kwargs": {"null": null}

when instead it should submit

"run_task_kwargs": null

Steps To Reproduce

1. Start an ECS agent
2. Navigate to "Run" screen on any flow page
3. Set the appropriate labels for the ECS agent
4. provide a valid task_role_arn
5. Run -- you'll see the above incorrect run task dict in the backend, and the run will error out with the error above

Browsers Tested:

• Chrome
• Firefox
• Safari
• Edge
• IE

---

Notes

The UI should submit null if no values are provided for run_task_kwargs cc: @kmoonwright

[znicholasbrown]

@cicdw since run_task_kwargs is a dict, an empty dict is valid as well, correct? #1029 fixes the issue where the dict was being sent as { "null": null }but a further update to the run component would be needed to filter out dict type fields that have no key/values and set them to null instead of an empty object/dict.

[cicdw]

Yea an empty dict should be fine 👍 (I haven't tested it yet though so we'll want to confirm)