Support for Retrying HTTP Service Calls in IBM App Connect
In this blog, we will learn how to support retrying HTTP Service calls in IBM App Connect.
IBM App Connect Enterprise (ACE) has introduced a new Retry capability for the HTTPRequest node in version 13.0.3.0. This enhancement allows the node to automatically retry failed HTTP requests under specific conditions, improving the reliability of API calls.
HTTPRequest Node Retry Mechanism:
The HTTPRequest node’s Retry properties control the ability of the node to automatically resend failed requests under certain conditions. This ensures that transient errors do not disrupt API interactions. The retry mechanism only propagates the result downstream when a response that does not match the configured error conditions is received from the endpoint.
- Retry mechanism: A drop-down menu provides options such as No retry and Short retry. By default, the setting is No retry. When Short retry is selected, additional parameters such as retry interval and conditions can be configured.
- Retry threshold: Specifies how many times the node should attempt a retry when it encounters the failure condition. By default, the value is 1.
- Short retry interval (seconds): Specifies how long the node should wait between retry attempts. By default, the waiting period for this property is 5 seconds.
- Retry condition: This allows you to specify the list of conditions that will trigger a retry. The list can include HTTP status codes (like 429), POSIX error codes on Unix systems, or Windows Sockets error codes (e.g., ECONNREFUSED).
You can customize this list through the Edit Condition option, which offers predefined choices of HTTP and socket error codes. If the error encountered matches one of these entries, a retry will be attempted. You can also restore the default set of error conditions if needed.
Working of Flow:
The diagram below shows the working of the HTTPRequest node’s Retry.
This is the main flow where I am using the HTTPRequest node and setting the properties for the Retry mechanism in the compute node.
This is the backend flow, which I am calling in main flow, and added the logic for sending status code 503 sometimes. When it gets 503 status code then it will retry again after a specified interval.
The new retry capability in the HTTPRequest node of IBM App Connect Enterprise 13.0.3.0 significantly improves API reliability by automatically handling transient failures. By leveraging this feature, integration developers can build more fault-tolerant, resilient, and efficient API interactions.
