Details
-
Improvement
-
Resolution: Fixed
-
Major
-
None
-
None
Description
The decision on when an operation is a failure is not trivial, and is hard to express on a general basis.
As such a failure policy framework should be established for use with the clients to determine when an operation is a failure. It should be possible to inject a failure policy when starting an operation.
A such policy could be "fail on identify timeout", "continue if at least one pillar is available" etc.
Attachments
Issue Links
- related to
-
BITMAG-655 Idempotent system
- Resolved