When Unity detects a request that failed on one system but succeeded on another, the session and tables with the failed request become interrupted. Lack of spool or perm space, missing user grants, and missing database objects can cause a request to fail.
The interrupted state means that Unity can still automatically recover the session and tables from the recovery log after the initial problem is resolved on that system.
When there is an interrupted request, Unity raises an alert to inform the DBA. The DBA can then fix the problem identified on the failing system. For example, the DBA can add more spool or perm space, correct user permissions, and create or drop required objects.
It is important to respond to the alert immediately because any subsequent work depending on the work done by the initial interrupted session also becomes interrupted. These secondary or cascade interrupts are filtered in the Unity UI and at the unityadmin command line so you can resolve the initial session that caused the issue.
If ignored, an interrupt alert can escalate from a small isolated issue, involving a single session and a small set of tables, to a larger problem that can cause the entire system to become interrupted.