Hi everyone, I have a question on
prefect_databricks
for the best practice. My flow run submitted several databricks jobs. It seems like I can use either
prefect_databricks.flows.jobs_runs_submit_and_wait_for_completion
or
prefect_databricks.jobs.jobs_runs_submit
.
I want to achieve the following goals in my use case:
1. we are able to get databricks job link in the logs,
2. we can easily set upstream dependencies,
3. we can get job state and set state handlers for error and retry
I tried both `jobs_runs_submit`and
jobs_runs_submit_and_wait_for_completion
, and feel like it's best to write our own wrap for
jobs_runs_submit
in order to achieve our goals above. Does that sound right? Any suggestions?
Thank you in advance!