As Carlton says, there was a 1 hour 16 minute gap between blocks 596166 and 596167. These kind of gaps have happened before and are not only possible but entirely expected given how mining works. Given that the hashrate graphs you see online are all estimated based on difficultly and the rate of blocks being mined, these graphs misinterpreted this gap as a drop in hashrate, when in actual fact it was a perfectly normal delay in blocks being mined.

In the hour before this gap, 10 blocks had been mined, which is way above average. Then we had a period which is way below average. If you look at the block production over several hours or days, it all averages out. Nothing happened except entirely normal random fluctuations, and people who don't understand how bitcoin works jumped on it as being some big news story.

It reminds me a lot of people who get panicked when they see a transaction of 10,000 bitcoins and are convinced the market is about to dump, and every time it turns out to be a complete non-story and just some exchange just moving their cold storage around.