Our APM reports a spike in latency at 4:04-4:05 pm eastern that led to a greater than normal number of timeout errors. Response times returned to normal, and our team investigated further. Our host confirmed the database was automatically restarted due to an underlying AWS EC2 incident.