In a production system where agents operate over long time horizons (minutes to hours), how would you manage state persistence and recovery?

Agents generate a lot of "internal thought" and tool logs. How do you keep the context window from filling up with irrelevant logs during a long multi-step task?

Why would you index sentences but provide the surrounding paragraph as context?

When would you use a "Manager Agent" to delegate tasks to "Worker Agents" rather than having one single agent do everything?