This feature was introduced in 24.1. It was designed to identify certain error message content and ask Automation manager to retry the task:

• a definable number of times
• with a specified retry interval

For example,  retry this task 3 times every 15 minutes before finally putting the task into error.

The challenge this solves is that each integration, particularly where middleware is involved, gives different and often quite technical error messages. If you know the error is temporary, usually due to a connection issue, you will usually want to retry the task.

The task error retry allows you to specify some or all of the words in the error message and based on finding those words, the task will retry.

Below we will work through an example based on a connection error which we would like Automation manager to retry.  We are going to take the highlighted text from within the error message and ask Automation manager to retry the task when that text is found.

The task error retry can be found in the Admin Tools menu in Automation manager. 

To create a new error retry press New.  

In the dialog box enter the text you want to look for within the error message. 

It's important to note that, as parts of the error message can change, such as a date and time stamp, you don't want to look for the whole message, but rather identify a piece of the message that will always appear. As you can see below we have only set Error retry pattern as NiHLGetNodeAddr: hostname cached which is the highlighted part of our example message.

Error retry pattern - set here the part of the error message that would trigger a retry.

Number of retries - determines how many times Automation manager will retry before putting the task into error.

Refresh interval - set the time in minutes between each retry.

Task Statuses - this is optional in that you may only want to retry when the task is running. Leave this empty to cover all statuses.

The Action log will record each time a retry is made.