A cloud-native SIEM solution that provides intelligent security analytics and threat detection across systems
The first check is whether the connector has completed its configured ingestion-cycle-days period, because ABAPUserDetails is intentionally collected progressively across that cycle rather than necessarily appearing in full during each poll.
Check the user-master collection cycle
The default ingestion-cycle-days value is 7 days, with supported values from 1 through 14 days. It controls how long the connector takes to ingest the complete User Master dataset, including all users and roles; changes to User Master data are handled separately.^1^
- Establish when this SAP client was connected or when the Data Collector configuration was last redeployed.
- Wait for one complete configured cycle before comparing SAP and Sentinel totals.
- Check whether records continue arriving:
ABAPUserDetails
| where TimeGenerated > ago(14d)
| summarize Records=count() by bin(TimeGenerated, 1d)
| order by TimeGenerated asc
A steadily increasing population during the cycle indicates phased collection, not necessarily user-type filtering.
Verify the effective Integration Suite settings
In the value-mapping artifact used by the Data Collector iFlow, inspect both:
- the
globalagency; and - the agency whose name exactly matches this SAP system’s RFC destination.
System-specific mappings can override global behavior. Parameters not present in the artifact retain their defaults.^1^
Confirm these effective values:
| Parameter | Required check |
|---|---|
collect-user-master-data-users |
Must be true; its default is true. |
ingestion-cycle-days |
Record the effective value, from 1 to 14; default is 7. |
After changing a value mapping, save and deploy the artifact. For a controlled diagnostic, temporarily setting ingestion-cycle-days to 1 can make the full cycle complete sooner, but monitor SAP and Integration Suite load while testing.
The published customization parameters do not document a selector that intentionally limits ABAPUserDetails to technical users. User and authorization details are supported sources for the agentless connector.^2^
If a complete cycle still produces only a subset
Because the source records and broad SAP authorization have already been checked, inspect execution completeness rather than repeating those checks:
- In SAP Integration Suite, open Monitor → Integrations.
- Locate Data Collector executions for the affected RFC destination.
- Inspect the Message Processing Log (MPL) for failed, retried, long-running, or abandoned user-master reads.
- If no useful error appears, temporarily enable trace logging and reproduce one collection run.
- Deploy the Prerequisite checker iFlow against the same RFC destination. Run it for up to 24 hours at one-minute frequency, then inspect its MPL attachments and per-check runtime measurements—particularly the user master read.^3^
The agentless poller enforces timeouts when calling the Data Collector iFlow; exceeding them can cause partial ingestion or retries. The prerequisite checker is the documented method for identifying a slow RFC destination or user-master operation.^3^
Information needed if the issue persists
Provide these items in a Microsoft support case, with secrets and personal data removed:
- Connector onboarding or last-redeployment timestamp.
- Effective
ingestion-cycle-daysandcollect-user-master-data-usersvalues from both global and RFC-specific mappings. - Expected SAP user count and current distinct
ABAPUserDetailscount after a complete cycle. - Sanitized examples of one ingested technical user and one missing dialog user.
- Data Collector MPL status, duration, correlation/message ID, attachments, and trace output.
- Prerequisite checker results and user-master-read runtime.
- Whether ingestion grows each day, stops at a repeatable count, or varies between cycles.
- SAP product/release, client number, RFC destination name, and deployed Sentinel Integration Suite package version.
These details distinguish normal cycle-based ingestion from iFlow timeout, batching, or package-level selection behavior.
References