Details
-
Task
-
Resolution: Unresolved
-
L3 - Default
-
None
-
None
-
3
-
M
Description
As part of importing Zeebe data, we deserialize exported Zeebe records and convert them to Optimize data (similar to what we do for fetched engine records). Historically, Optimize has used the same terminology that the engine used. In the case of Zeebe, some of this terminology differs, and it results in bug-prone and confusing conversion code. For example, the key and ID of the process is reversed.
We should instead make this conversion to the common terminology during deserialization of the zeebe record into an Optimize DTO. This would keep the confusing mapping in a single place, i.e. deserialising the bpmnProcessId to the processDefinitionKey field rather than the processDefinitionKey field of the Zeebe record, which is actually the ID in an Optimize sense. There might be other consistencies that could be handled this way too.
ATs:
- After Zeebe records are deserialized, we should use a DTO with consistent naming/terminology as the rest of Optimize for elements of a process instance