At 03:14 PM we lost one of our Search engine nodes used by our Logistics service and immediately started the recovery process which was finished at 03:33 PM with full recovery of our service.But at 05:27 PM UTC-3 we lost another node of the same cluster. We started the recovery process again until 05:51 PM UTC-3 when the service was fully recovered.During the recovery period we experienced elevated error rates in our platform. This incident had a partial impact on sales, affecting mainly our Logistics service which is part of the critical path of Checkout.The service is now operating normally, we have identified the cause of the issue, and our engineering team is testing measures to prevent it.