You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Implementing this versioning system makes me feel that this stack is slightly over-convoluted as it stands- especially having a whole separate deployment for each job.
A simpler way to do it would be to have a single deployment for each entry point (conceivably in a single prefect.yaml), and store recipe configuration stuff in S3 or another external storage. This is a significant refactor of the approach I've been using so far, but I think it's a preferable separation of concerns.
The text was updated successfully, but these errors were encountered:
First steps in a recent commit- we CAN load a recipe directory directly from S3 via a new entry point.
That entry point could use some massaging.
Next step would be to create a sous-chef prefect deployment file that's JUST sous-chef.
Then, we might want a script to make uploading recipes to s3 more straightforward.
Implementing this versioning system makes me feel that this stack is slightly over-convoluted as it stands- especially having a whole separate deployment for each job.
A simpler way to do it would be to have a single deployment for each entry point (conceivably in a single prefect.yaml), and store recipe configuration stuff in S3 or another external storage. This is a significant refactor of the approach I've been using so far, but I think it's a preferable separation of concerns.
The text was updated successfully, but these errors were encountered: