If batch job is possible and you need to process data I would use probably:

azure event hub from (events after previous job run) > databricks job process as dataframe > save df to delta lake

no streaming or capturing needed in that case