Preventing continued distribution of a software update that is causing problems in computers is a challenging problem, particularly where the update causes a catastrophic failure such that the problem cannot be reported by the computer since the computer has been completely disabled. To manage this problem, when an update is delivered for installation, it first installs a program and configures it to execute at a specified reporting time. When that time is reached, the program sends a positive operations notification to the update server indicating that the program is okay or sends a notification that the program is okay so far, but the user is now shutting down the computer. The number of notifications received is tracked by the system in comparison to the number of software updates sent. If the notifications received are smaller than expected, the update may be causing catastrophic failures in the computers, preventing the sending of the notification, and thus the system can take corrective action to manage this problem early on, before more problematic updates are sent out.