Since yesterday, we have not been able to refresh our Power BI dataset with the Autotask source from Invantive. When checking the Monitoring Bridge Online, I see the following error:
The database ‘ACME: acme-autotask’ could not be opened.
When attempting to open the database, I receive the following message:
"{"error":{"code":"itgenclr003","message":"The database \u0027ACME: acme\u0027 could not be opened. (itgenclr003, a39bd606-66d2-41d8-a73f-f33b9c3b9dda)","target":null,"details":[{"code":"a39bd606-66d2-41d8-a73f-f33b9c3b9dda","target":null,"message":"Unique ID"}],"innererror":null}}"
This means that the metadata can be replaced at any moment by the Autotask platform partner datto, even when the Invantive platform does not change.
In general new tables and fields are automatically reflected in the data model. The new names can sometimes need some love in terms of better naming, but it should keep working. This allows Invantive SQL to provide all available data elements.
The automatic mapping however bears the risk that the mapping does not handle all new constructs automatically and may raise an error. In this case, an error is raised by the automatic mapping.
An analyst of Invantive will look into the issue. Once the cause has been established, an adapted version of the automatic mapping will be implemented and deployed.
This can take a few working days, since the improved mapping must be incorporated in a release with major changes for reduced issues on Exact Online time-outs and task scheduling. This release is scheduled for deployment by next Monday, but planning may vary depending on the tests being executed.
This answer may be extended by describing a workaround in between time.
An analysis of Autotask has been completed. The provided dynamic datamodel does not meet the OpenAPI specifications in the last release. A future release of Autotask may fix this.
A workaround has been implemented which makes the datamodel comply to the standard again.
A release will be deployed after regression tests. This typically takes a few hours, excluding unexpected conditions.
The topic will be updated once deployed in production.