A streaming platform staffed every live event with a full incident bridge because concurrency spikes were unpredictable and expensive to get wrong. The SRE Orchestrator made the bridge an exception rather than the default.