System Administration, Performance, and Maintenance
Domain 8 tests the operational discipline that keeps the platform stable: housekeeping, tuning, backup, and change control. Performance problems are usually data growth, configuration drift, or resource exhaustion.
Maintenance
- Housekeeping/purge: archive or delete old process data on a schedule; uncontrolled growth slows queries and backups.
- Database: monitor table and index growth, run statistics, and size the transaction log.
- Logs: rotate and retain according to policy; logs on a full disk stop processing.
- Certificates: track expiry for SSL, signing, and partner trust.
Performance
Tune JVM heap and garbage collection, thread pools, adapter container count, and perimeter server capacity. Measure before changing: identify whether latency is in the process, adapter, database, or network. Scale adapters horizontally when polling or connection count is the bottleneck.
Backup and recovery
Back up the database and shared file system consistently; a database backup without deployed artifacts and properties is incomplete. Test restore, not just backup. Patch in a staging environment and keep customer_overrides.properties under version control.
Worked example
Process throughput degrades over months. The database has millions of completed process rows and the purge job was disabled after an upgrade. Re-enable housekeeping with a retention window, archive required audit data, and run during a low-volume window. Reasoning: the application code did not change; data volume did.
Traps
- Purging data still needed for audit or dispute resolution.
- Tuning heap without checking a database lock or slow query.
- Backing up only the database and losing configuration.
- Applying fix packs directly to production.
- Ignoring disk space on the shared file system.