Provide an option to resume when failed.

[Naveen-Bhat]

When migrating a medium to large project, if we get an error at the end of the process, its really frustrating to start from the beginning. Wherever possible, please provide an option to resume from where it left.

Ex: Network error. I believe if local state is maintained then its not necessary to fetch the data again from the source and a significant time can be saved.

For few errors, its better to do the validation prior rather than later so that the process doesn't begin until these errors are resolved. Ex:

[10:53:19 DBG] [16.0.9] 12931 Work Items loaded
[10:53:19 INF] [16.0.9] Replay all revisions of 12931 work items?
[10:53:19 INF] [16.0.9] Validating::Check all Target Work Items have the RefectedWorkItemId field
Initializing expensive ProjectData from WorkItemStore...
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Epic contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Feature contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: User Story contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Task contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Bug contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Test Plan contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Test Suite contains Custom.ReflectedWorkItemId
[10:53:19 DBG] [16.0.9] ValidatingRequiredField: Test Case contains Custom.ReflectedWorkItemId
[10:53:19 WRN] [16.0.9] ValidatingRequiredField: Shared Steps does not contain Custom.ReflectedWorkItemId
[10:53:19 WRN] [16.0.9] ValidatingRequiredField: Shared Parameter does not contain Custom.ReflectedWorkItemId
[10:53:19 ERR] [16.0.9] Not all work items in scope contain a valid ReflectedWorkItemId Field!
Microsoft.TeamFoundation.WorkItemTracking.Client.InvalidFieldValueException: Not all work items in scope contain a valid ReflectedWorkItemId Field!

[MrHinsh]

The whole process once started is resumable.

However loading the data to process, and analysing it for faults is not resumable, and I don't see a way to save and load it.

---

Id suggest running checks on a smaller subset of data... like "last 30 days".