Reading the post mortem, I'm guessing it's more like it took them 30 minutes to figure out what was happening and then they rolled back the deploy and it was fi

Reading the post mortem, I'm guessing it's more like it took them 30 minutes to figure out what was happening and then they rolled back the deploy and it was fixed. Edit: Reading further down, they actually admit that it was just a roll back > At 12:44 PDT, the team discovered the root cause, the configuration change was promptly reverted, and the affected GFEs ceased their restarts