Share all details about your problem, including any error messages you may have received.
Our team is about to integrate Workday HR system using Workday OOTB connector to replace the current delimited Workday HR connector. Is there any guidance on high level cutover plan and any recommendation on the convention?
Parallel Integration
Deploy the Workday connector in parallel with your delimited one. Use a different application name (e.g. Workday-API) and configure correlation on a stable key like employeeID or WorkdayID. Avoid relying on email unless you’re certain it’s immutable in Workday.
Schema Alignment & Normalization
Map attributes carefully — especially booleans (e.g. isManager), dates (hire/termination), and organizational units. Workday often returns wrapped or nested values, so expect to use XPath or response parsing to flatten values.
Identity Correlation Validation
Run the connector in read-only mode and execute a full identity refresh. Pay attention to:
Unmatched identities
Duplicate creation
Inactive user behavior
Manager hierarchy integrityCorrelation issues typically stem from inconsistent casing, nulls in primary keys, or slight data differences in transforms.
Lifecycle Event & Trigger Testing
If you use joiner/mover/leaver logic or RapidSetup events, test them thoroughly. In some cases, Workday returns “future-dated” status changes — make sure your logic accounts for that (e.g. future terminations or future hires).
Switch Authoritative Source
Once validated, disable aggregation on the delimited connector, flip the authoritative source flag to the Workday API app, and re-aggregate.
Post-Cutover Monitoring
Watch identity refresh logs and source link mappings. Consider using a custom report or quick rule to catch unexpected population changes during the first few runs.
Additional Tips
XPath tuning matters. The Workday connector gives you flexibility with XPath expressions. Use the SailPoint Workday IDE tool or trace-level logging to refine them.
Future-dated logic: Workday may show a user as Active today even if they’re already terminated in the future. Handle this via transforms or filters if necessary.
Avoid full refreshes for deltas. Once live, schedule incremental fetches based on the Workday Last Modified field rather than full imports — reduces load and speeds things up.
Naming & Conventions
Keep the old application (e.g. Workday-Delimited) around but disabled — don’t delete it until after multiple successful runs.
Document your correlation logic and XPath mappings. These tend to be tribal knowledge that gets lost over time.
Hope this helps! Let me know if you have any questions.