1. Scheduled agents should be monitored to ensure they are running successfully on servers. A central log database can be used to store status documents from agents and alert if reports are overdue.
2. Logging is important for scheduled agents but the NotesLog model can result in large log files. It is better to develop a "ticker-tape" style log facility that stores each log entry in a small document.
3. When designing scheduled agents, consider how to handle processing large amounts of data within the limited runtime. The agent may need to run multiple times and checkpoint processing to avoid timeouts.