Calling unreliable downstream systems
We are calling some unreliable downstream SOAP web services from HPOO.
They quite often take longer than 30 minutes to complete or have intermittent failures.
What is the best practice in HPOO for retrying failed requests? We are thinking of things like increasing timeouts, adding delays between requests, etc.
Also are there any existing built-in or out of the box components we can use for this?
Which version of OO are you using- 9.x or 10.x ? And how exactly are you calling the web services- via API or through the OOB content ? In fact, the workarounds you've mentioned are viable solutions in situations like this one you just have to choose which one will suite you best, based on the common outcome of the calls.
We are currently using 9.x and in the process of upgrading to 10.x, will this upgrade provide any additional benefit when calling web services?
The flows mainly seem to use the Web Service Wizard to call the web service via RAS and the WebServiceInvoker.jar.
I'm wondering if the best approach is to create a new configurable flow that takes two lists, one for timeout and another for retry waiting time. Then the component can be invoked to call the web service with increasing timeouts e.g. 1,2,5,10,30 mins etc and also with pauses between each call to ensure that the end system is not unecessarily overloaded. Does this seem reasonable or there is something built in we can use?