Revisiting Multi-Agent Asynchronous Online Optimization with Delays: the Strongly Convex Case