We encountered an issue causing some requests to take longer than expected. We pushed a fix and scaled accordingly. Everything should be back to normal.