Earlier in the morning, our GPU Provider Paperspace announced that they had a downtime. We have a GPU instance responsible for Translation. They took quite some time to solve it, more than 15 hours. Over the day, on the 1st of December, we switched to the fallback solution. Now everything is back to normal, but for the whole day, you might have experienced fewer than usual articles with translated fields. We are going to reprocess everything that was lost, and we suggest on the 2nd of December to redo the queries for the 1st of December to make sure you get the data. Sorry for the inconvenience.