Currently the raw input/output files and logs are available.
It would be nicer for users to have a dashboard which provided some filtering and statistics.
As we control the whole pipeline, emitting additional information regarding stages run and failure/completion would be relatively trivial.
Storing the metadata in a database (tools-db) and pointing to the files via the existing files api would keep things speedy (the directory structure grows daily).
Extending the existing proxy image would allow the existing API to be kept separate got now, potentially being merged down the line.
The existing code should be lightly refactored so the coordination job (Steps wrapper) is more generic.
Currently the raw input/output files and logs are available.
It would be nicer for users to have a dashboard which provided some filtering and statistics.
As we control the whole pipeline, emitting additional information regarding stages run and failure/completion would be relatively trivial.
Storing the metadata in a database (tools-db) and pointing to the files via the existing files api would keep things speedy (the directory structure grows daily).
Extending the existing proxy image would allow the existing API to be kept separate got now, potentially being merged down the line.
The existing code should be lightly refactored so the coordination job (Steps wrapper) is more generic.