Controlling Imported Data and Hosting Boundaries
Key Takeaways
For supported connectors, use source filters or query projections to limit which rows and attributes are sent to SailPoint.
Account schema and identity attribute mapping serve separate purposes. Omitting an identity mapping can leave the raw source-account field in the cloud.
Core tenant hosting region, local VA placement, and ancillary processing destinations are distinct.
Connector credential encryption controls key location. It does not keep all aggregated identity data inside the VA network.
Decide what can leave the source
Map the requested governance outcome to the minimum accounts, attributes, entitlements and memberships it needs. Three points matter. First, the connector’s filter or query can limit the returned dataset. Second, SailPoint’s account schema specifies which source-account fields it will store and how they are represented. Third, an identity profile can map some of that data into tenant identity attributes. Leave attributes out of the connector query or feed if they should not be sent. A missing identity mapping only shows that the selected profile does not populate that identity attribute from the value. A collected field can still be present in the source account record.
| Requirement | Control to examine |
|---|---|
| Limit which private directory users are loaded | User Search DN and a supported LDAP filter for the selected Active Directory connector |
| Limit directory memberships independently of users | Group Membership Search DN and membership filter |
| Limit database fields and rows | Approved database view or explicit SQL columns and WHERE predicate |
| Limit a flat file feed | Export only approved schema columns and rows, and verify the import format |
| Validate what the tenant actually retained | Inspect the source account and entitlement data as well as relevant identity attributes |
Suppose a personnel database will populate an employee population. The business needs a stable account identifier, last name, work email, and a department code for role criteria. It has no requirement for the employee’s salary. Use an authorized view or query that selects the approved fields while excluding salary. Align the account schema with the returned column names and confirm that sample account records have the right fields. A particular access policy may use only department, while the identity profile still contains the required mappings. The source query is where the engineer can enforce the data boundary for this feed.
If a team narrows a filter after earlier collections, inspect the existing tenant inventory and the consequences of the new scope. Filtering future input cannot prove that records collected earlier have been removed. A full aggregation with account deletion enabled handles missing accounts according to documented safeguards. A tenant account removal is a different operation from deleting the account on the managed system. Plan the handling of historical copies, derived fields and reports from available product controls and the organization’s policy.
Document the expected result as a short field inventory for each feed. Compare the connector result, stored account fields and mapped identity attributes on the same sample identity. When an excluded value still appears, this comparison identifies whether it came from another source, an earlier retained record or a derived identity mapping.
Identify hosting and credential boundaries
The tenant hosting region concerns the hosted service. SailPoint’s subprocessor disclosure describes Amazon Web Services hosting for core Identity Security Cloud services in the region the customer elects from available options at the start. Confirm the actual tenant, contracted product and its approved hosting location. A private VA deployment location shows where connector operations run. It does not specify where the SaaS tenant stores the records it receives. Check relevant product-specific and subprocessor disclosures separately when a case involves telemetry, an integration, support or an add-on with other processing locations.
A VA-based connector can encrypt credentials using a cluster public key in the cloud and decrypt them with the private key on the appliance. That model describes certain connector secrets. Account records intentionally aggregated into the tenant are a different data flow. SaaS connector key handling can also differ because it has no customer appliance holding the private key. Trace each category of information to the actual system and control that applies.
In a design review, an architect argues that a VA in Germany keeps all connector activity and all data in Germany. Verify which fields the connector will send and which cloud tenant receives them. Check the customer’s selected cloud region and applicable processing destinations rather than inferring them from appliance placement. If the team is asked to change where the data is hosted, changing an LDAP filter alone will only change what the connector queries. That setting does not configure the hosted tenant.
See SailPoint’s account schema, JDBC query settings, AD search settings, hosting and subprocessors, and credential encryption design.
A JDBC source currently selects salary and work email. Work email is needed, but salary must not be sent to the hosted tenant on future runs. What should the engineer change?
Remove only the salary mapping from the identity profile while keeping the current query
Install the VA closer to the database
Remove salary from the connector’s approved query or view and check the returned schema and payload
Change the tenant administrator’s search privileges
A private VA runs in one country. What is the best evidence for where the core hosted Identity Security Cloud tenant stores aggregated data?
The tenant’s selected hosting region and the applicable product hosting documentation
The physical address of the VA’s server alone
The region used by the target database only
The fact that the VA keeps a private credential-decryption key
Sections you finish are checked off in the contents.