Choosing Configuration, Transforms, or Rules
Key Takeaways
Start with built-in mapping, policy, and operation settings for supported direct behavior.
Use a transform for a supported calculation and choose input types for the actual context.
Choose a supported rule only where the declared hook and maintainability justify custom logic.
Test shared dependencies and effective output before promoting a changed rule or transform.
Decide where the result is consumed
Determine whether the customer needs a mapped value, an existing value in a provisioning policy, complex logic at a specific supported hook, or an integration that performs more than a data calculation. If a single source attribute is the exact value needed, configure the built-in mapping. If primitives can produce the desired value and can be understood by the next operator, use a transform. If there is a legitimate gap and a supported hook supplies the inputs, evaluate a rule. For a separate action in an external system, consider the correct integration or workflow rather than turning an attribute calculation into an external call.
| Scenario | First fit to evaluate | Why it fits |
|---|---|---|
| A clean identifier maps directly to an identity field | Direct mapping | No calculation needs to be deployed |
| Several profiles need the same maintained code table | Shared Lookup and Reference transforms | Keeps one declared mapping that dependent profiles use |
| Create policy needs a value already promoted to the identity | Provisioning policy using the supported identity input | Reads available identity data in the account-creation context |
| A required target username needs custom supported generation and uniqueness behavior | First evaluate Username Generator, then a supported field rule if needed | The native account context is the intended consumer |
| External service must be called to open a ticket | Choose the appropriate service desk integration or workflow | The required outcome is an external action with its own evidence |
| Cloud calculation requires a custom supported algorithm the available configuration cannot express maintainably | A matching supported cloud rule | Restricts execution and review to the actual hook |
Work a choice across identity and provisioning
Two profiles read a region code from authoritative HR, and one application needs the normalized region when it creates its accounts. Use a shared declarative transform to derive the region in the identity mapping where the source data is available. Verify mapped region values from both profiles. For creation, refer to the existing identity attribute through the field's supported identityAttribute input. SailPoint's provisioning policy guide documents that accountAttribute does not work for this use. Test the output of the identity calculation before validating the Create Account field and its native effect. A configuration that works inside an identity profile can still be incorrect at a provisioning hook if the input shape and dependencies differ.
Next, suppose the target requires a unique account identifier. Test the documented Username Generator capability and its connector constraints first. If a custom field generator is required, verify the appropriate rule, supported uniqueness checks, and native limits. Do not derive a changing unique identifier by repeatedly re-running an Identity Attribute rule. SailPoint warns that identity values can recalculate and multithreaded identity calculation makes this an unsuitable place to rely on a uniqueness check. Verify the actual target generator and collision handling with a controlled create operation using synthetic data.
Choose tests and promotion scope
For a shared transform, list dependent profiles and policies before editing. A single named reference can reduce duplicated configuration, and changing its output affects its consumers. Test representative values from all affected sources, including null, unrecognized codes, and malformed input. A transform can use a rule through the supported rule transform and the appropriate Transform rule type, but that code is still governed by the cloud rule process. Pick this combination for a supported, maintainable need after testing native transform operations.
Keep the approved rule or transform definition, its dependencies, and the expected test cases in version control. Review what can be promoted and which tenant-specific references must be configured, then prove outcome and recovery on a scoped sandbox sample. Use the supported deployment mechanism for the artifact type. For an approved cloud rule already installed elsewhere, use the applicable protected configuration import procedure. An unapproved artifact belongs on its rule review path. After deployment, compare expected outcomes and actual mapped or target values, and retain the previous approved configuration where available. Use the information to revert the specific change if the acceptance conditions fail.
See SailPoint’s transforms and rule selection, reference operations, provisioning policy data, rule deployment guidance, and Identity Attribute restrictions.
Two identity profiles use the same normalization table. The table must remain consistent over time. Which supported design best serves that requirement?
Use separate copied tables and update whichever one is visible first.
Use one named Lookup transform with references and test each consuming profile before updating the table.
Create an unrelated connector rule in each target application.
Add two Before Provisioning rules to rewrite the records after aggregation.
A Create Account configuration must use an already normalized region that is on the identity. Which input is appropriate for that field?
An Account Attribute transform that tries to find a newly created target account.
A connector read without checking its operation support.
The supported identityAttribute input for the mapped region in the Create Account policy.
The Previous Value argument for an unrelated manager rule.
Sections you finish are checked off in the contents.
You've completed this section
Continue exploring other exams