Asynchronous request administration
Performance considerations
The architecture for asynchronous requests is designed to support a large number of asynchronous requests on a daily basis. However, it is not designed to support every request being executed asynchronously. Guidewire recommends that you execute calls asynchronously only when necessary.
The AsyncApiRequest data model entity
Information about asynchronous Cloud API requests are stored in the
AsyncApiRequest entity. You cannot extend this entity.
The PurgeAsyncAPIRequests batch process
The PurgeAsyncAPIRequests batch process removes information about asynchronous Cloud
API requests from the AsyncAPIRequest database table that are more
than X days old. In the base configuration:
- The value of X is 7.
- The batch process is scheduled to run once a day at 3:30 am.
You can configure the value of X by manually adding an
AsynchApiRequestPurgeDaysOld configuration parameter to
config.xml. For example, if you wanted request information to
be purged when it is 6 days old, you would add the following to
config.xml:
<param name="AsyncApiRequestPurgeDaysOld" value="6"/>You can also configure when the batch process is scheduled to run. For more information, see the Administration Guide.
Limitations on user session access
If a particular endpoint requires a session, or tries to create one, it cannot be used with the async header. The Cloud API endpoints never do this. This limitation applies only to a custom endpoint an insurer has built on the REST API Framework.
Any code that optionally looks at the user’s session to see if it is there will find that it is null when the request is made asynchronously.