Configuring Retry Functionality for Sending Responses
Integration Server uses a transactional publish when sending responses to IBM webMethods Cloud. The transactional publish includes a built-in mechanism to recover from intermittent network errors, which can help mitigate the impacts of timeout errors and connection leaks. Integration Server automatically retries any failed transactional publish.
Use the following server configuration parameters to configure the retry functionality for publishing responses to IBM webMethods Cloud.
watt.wmcloud.sendResponse.retryCount Specifies the maximum number of retry attempts that an on-premises
Integration Server makes when a failure occurs when sending a response to
IBM webMethods Cloud. The default value is 3. A value of 0 indicates that
Integration Server will not attempt any retries.
Integration Serverdoes not need to be restarted for changes to this parameter to take effect.
watt.wmcloud.sendResponse.retryDelay Specifies the number of milliseconds that an on-premises
Integration Server waits before making another retry attempt after a publish attempt fails. The default value is 1000 (i.e., 1 second).
Integration Serverdoes not need to be restarted for changes to this parameter to take effect.
If the maximum retry attempts have been made and publishing a response to the IBM webMethods Cloud still fails, the on-premises Integration Server writes an error to the server log. However, if a response payload exceeds the configured maximum buffer size, then Integration Server does not retry publishing the response to IBM webMethods Cloud but sends an exception to IBM webMethods Cloud.
Note:
Setting the server log facility 0021 Integration Cloud to the Info level or more verbose level may assist with debugging any publish failures.