Business Process Scheduling, Monitoring, and Troubleshooting
Domain 7 tests scheduling, monitoring, and diagnosis. You need to know where to look first and how to distinguish a process problem from an infrastructure problem.
Scheduling
The Scheduler service starts business processes on a defined calendar or cron-like schedule. Check time zone, start/end dates, blackout windows, and whether the schedule is enabled. Scheduled processes should be idempotent and should not overlap.
Monitoring
- Business Process Monitor: process status, current service, elapsed time, and errors.
- System Monitor: adapter and server health, queue depth, and resource signals.
- Alerts and logs: process logs, adapter logs, and system logs; correlate by process ID and document ID.
Troubleshooting order
Confirm the process is actually running, identify the current service, read the service error, check the adapter connection, then check the database and JVM. For hangs, take a thread dump and look for blocked threads, database locks, or a waiting adapter timeout.
Worked example
A file is not picked up. The schedule fired and the process is "waiting" at the file system adapter. The pickup directory is correct, but the file mask is case-sensitive and the file arrived as .TXT. Fix the mask or standardize naming. Reasoning: the process is healthy; the selection criterion is wrong.
Traps
- Restarting the server before reading the service error.
- Ignoring time zone differences in schedules.
- Treating an adapter warning as a process failure.
- Overlapping schedules causing duplicate documents.
- Looking only at process logs when the adapter never connected.