redundant job_id information

[hoiting]

It seems like to trigger a job, one can do
databricks jobs run_now <job_id>
however, if one also wants to provide a json, the job_id info needs to be provided again in --json or the command will fail

databricks jobs run_now <job_id> --json '{"python_params": [...]}' <-- this will fail

databricks jobs run_now <job_id> --json '{"job_id": <job_id>. "python_params": [...]}' <-- this will run fine

The same information is provided twice. It is a bit inconvenient and unclear which one gets used in the end.

[andrewnester]

The command can either be run as databricks jobs run_now <job_id> or as databricks jobs run_now --json '{"job_id": <job_id>. "python_params": [...]} (note, positional argument is not needed)

[hoiting]

Thanks. This behaviour is not consistent through out the cli though. For example, databricks serving-endpoints update-permissions <endpoint_id> --json <json_payload>, it seems like one has to provide both for it to run. Simply having the id in json but not in the positional argument won't work.

[andrewnester]

CLI commands behaviour was largely improved in this PR to avoid inconsistencies #878 and released in latest version of CLI.