Hi Community, We are facing an SSL issue while calling an external SAP API from our Mendix application. Environment Details: Mendix Version: 9.18.3 Hosting: Mendix Cloud Integration: REST API call to SAP Occurrence: Intermittent Error Details: Error message: SSL peers shutdown incorrectly Exception: java.io.EOFException Issue Description:The error occurs intermittently during the API call from Mendix to SAP. When the issue happens, the API fails to fetch data from SAP. At other times, the same API works as expected. We have verified the SSL certificate on the SAP side, and it is valid until October 10, 2026, so certificate expiry does not appear to be the cause. Questions / Help Needed: What are the common root causes of SSL peers shutdown incorrectly with java.io.EOFException in Mendix Cloud? Could this be caused by: TLS version or cipher suite mismatch? SAP server closing the connection without a proper SSL/TLS shutdown? Network, firewall, or load balancer timeouts? Are there any Mendix Cloud or JVM-level settings (TLS, JDK, HTTP client configuration) that should be reviewed? Has anyone experienced similar intermittent SSL issues with SAP integrations on Mendix Cloud, and how was it resolved? Any guidance, troubleshooting steps, or best practices would be greatly appreciated. Thanks in advance.
asked
Dharshana A
2 answers
0
Hi Dharshana,
This error relates to a disrupted connection. Some gateway/firewall/server is killing the connection (unexpected). Is the call a long-running one? Maybe you can change your call to an async setup?
Good luck!
Kind regards,
Johan
answered
Johan Flikweert
0
You should check the logs to clearly see how long the request runs and where the connection drops (endpoint, duration, timestamps). Then, if there is a load balancer, firewall, or proxy in between, ask to review idle timeout and connection reset settings. On the application side, adding a retry mechanism (for example, 2–3 retries with a short backoff) for idempotent requests like GET often helps. Also make sure the timeout values are realistic and review the keep-alivebehavior to avoid issues with reused connections. If the requests are long-running, switching from a synchronous call to an asynchronous approach and fetching the result later can significantly reduce the risk of intermittent connection drops.