A scheduled post fires. The request to publish it goes out. The network hangs. After 30 seconds, our client times out. We retry. The tweet publishes. Then the original request completes too — and a second, identical tweet goes out. That's a double-post. On a personal account it's embarrassing. On a client account at an agency it's a support ticket and a credibility hit. Either way, it's the single most visible failure mode in any posting system, and it's caused by one of the most common condi...