Symptom
An SAP Cloud Integration (CPI) iFlow fails intermittently with the following error in the Message Processing Log (MPL):
com.sap.it.rt.adapter.http.api.exception.HttpResponseException:
An internal server error occured: HTTP Request failed with error : Read timed out.
The MPL ID for the failed message is : <MPL_ID>
- The failure is not consistent — some runs succeed while others fail on the same iFlow.
- The iFlow sends a large number of sequential OData PATCH requests to SAP C4C (Sales Cloud).
- The error occurs on the CPI side, not on C4C.
Environment
SAP Cloud Integration (CPI / SAP Integration Suite)
Cause
The failure is caused by the Timeout (in min) parameter on the OData V2 Receiver Adapter being set too low relative to the response time of the target system under load.
Explanation:
- The default/configured value of Timeout (in min)= 1 corresponds to **60 seconds**.
- The iFlow sends a large number of sequential PATCH requests to C4C.
- Under high load, C4C can occasionally take longer than 60 seconds to respond to a single request.
- When C4C's response time exceeds the configured timeout, CPI terminates the HTTP connection and marks the entire message as failed — even if C4C completed processing of the record successfully on its side.
Why the behavior is inconsistent
- On runs where **all** C4C responses arrive within 60 seconds → the iFlow **succeeds**.
- On runs where **any single** PATCH call exceeds 60 seconds → CPI terminates the connection and the **entire message fails**.
This explains why the same iFlow can succeed and fail across different execution runs without any code changes.
Resolution
Immediate Fix — Increase Timeout (in min) on the OData V2 Adapter
1. Open the iFlow in the SAP Integration Suite designer.
2. Navigate to the **OData V2 Receiver Adapter** configuration.
3. Locate the **Timeout (in min)** parameter under the **Processing** tab.
4. Increase the value to accommodate slower C4C responses under load. Recommended: **2–3 minutes**.
5. Save, redeploy the iFlow, and monitor subsequent runs.
Additional Recommendations
1. Implement Error Handling / Retry Logic at the Individual Call Level
A single slow C4C response should not cause the entire message processing run (thousands of records) to fail. Consider implementing:
- **Exception subprocesses** around the PATCH call to catch `HttpResponseException` and retry the individual request.
- A **retry counter** to limit the number of attempts per record before escalating.
2. Long-Term — Migrate to OData `$batch` Processing
The root cause is amplified by the sequential, one-call-per-record design. With a high number of PATCH operations:
- Each call is an independent HTTP round-trip, increasing the cumulative probability of a timeout.
- `$batch` processing bundles multiple operations into a single HTTP request, dramatically reducing the number of network calls and the overall risk of timeouts.
Implementing `$batch` is the recommended long-term architectural improvement for iFlows with high-volume OData write operations.
See Also
Keywords
CPI, Cloud Integration, OData V2, TimeOut, Read timed out, HttpResponseException, PATCH, C4C, Cloud for Customer, SAP Integration Suite, MPL, sequential requests, iFlow, OData batch , KBA , LOD-HCI-PI-CON-OD , OData Adapters , Problem
SAP Knowledge Base Article - Public