Text this: Maximizing Service Reliability in Distributed Computing Systems with Random Node Failures