dstack
dstack copied to clipboard
[Feature]: Introduce run dependancies
Problem
Some runs may need to be run sequentially and depends on a previous run completing successfully.
Solution
Introduce a depends_on field, in which a run can be given another run ID in which it is dependant on. Once the previous run finishes, the current run can start.
This could be expanded to have different start reasons. Initially, starting a run when a previous run is successful will be the most common use case. If the previous run fails, so does the dependant run. However I could potentially see wanting to start a run if the previous job fails.
Workaround
Use of third party tools to coordinate dependant runs.
Would you like to help us implement this feature by sending a PR?
No