This is a story of a race against time! So hang on to your seats...

During a customer migration to a new attribution system, a huge project for AppsFlyer Engineering in 2018, we found ourselves facing a potential data loss catastrophe. It all started with the primal sin of a premature optimization made where we set the incorrect data retention timeframe for a database holding 65 billion records.

When we discovered this, with only one week to respond before the data is permanently erased, we channeled our Macgyver skills and got to work. During this session I'll describe the chain of events that brought us to the cliff's edge, the steps we took around the clock to save our data, and how we managed to forestall any data loss for our clients.