PROBLEM:
When zillions of systems are given an action to perform at roughly the same time, the systems check in with a crude “attempt, fail, retry” loop. This is a performance and execution issue.

Who effected:
All customers initiating actions with more systems than Apache/DB connections available.

Requirement:
Those connections should be “managed”, i.e., proactively queued, and the estimated checkins for each noted as to when the expected actions will be initiated.