Data Pipeline Health Monitor
Every target is checked independently for file mtime (OS-level modification time) and last data record (last Date row in the CSV), each against a per-frequency latency window.
All monitored files are declared in monitor.yaml with group, frequency (weekday / daily / weekly / monthly), latency override, and date column. Add or remove targets with one line.
Report is generated on x260 each trading day at 11:30 and synced to production. The JSON carries hostname + IP to identify the source machine.