As a person responsible for a live workflow, I want the history page and the editor to be one path, so that I can go from a list of failures to the thing that caused them without starting again.
Part of #4852 (Sandbox DevX epic). Phase 2.
Outcome
A design, then the work that follows from it. Nothing to build until the first part is done.
The lifecycle work designed the editor: which version is live, how to roll back, how to take it offline, how to change it safely. That is the right surface for changing a workflow. It is not where the day starts. Whoever looks after a live workflow opens the history page, filters to what failed overnight, and works from there. The editor's own history panel is a slice of the same data with none of the filtering.
So there are two surfaces for one job and no designed path between them. Going from a failure in the list to the version that produced it, to the fix, and back to retrying everything that failed the same way, currently means the person holding it together in their head.
Scope
- Work out what the person actually does between the two, and where each surface should hand over.
- Then open the pieces that come out of it.
Notes
Raised while reviewing whether the live-workflow design is finished. It is, for the editor. This is the part nobody has looked at.
As a person responsible for a live workflow, I want the history page and the editor to be one path, so that I can go from a list of failures to the thing that caused them without starting again.
Part of #4852 (Sandbox DevX epic). Phase 2.
Outcome
A design, then the work that follows from it. Nothing to build until the first part is done.
The lifecycle work designed the editor: which version is live, how to roll back, how to take it offline, how to change it safely. That is the right surface for changing a workflow. It is not where the day starts. Whoever looks after a live workflow opens the history page, filters to what failed overnight, and works from there. The editor's own history panel is a slice of the same data with none of the filtering.
So there are two surfaces for one job and no designed path between them. Going from a failure in the list to the version that produced it, to the fix, and back to retrying everything that failed the same way, currently means the person holding it together in their head.
Scope
Notes
Raised while reviewing whether the live-workflow design is finished. It is, for the editor. This is the part nobody has looked at.