Here is the response from our Product Team:
The description of the Create mode by the customer is correct. If they have the connector properly configured then https://<fhir-server-url>/Patient should return all patients, including those created by the connector. If they are indeed calling the Patient endpoint (no further query parameters) then a few things could be going on.
• They aren’t authenticated when they are issuing the get via postman. The customer isn’t providing the entire operation outcome, but authentication failed is a valid operation outcome.
• They are providing additional query parameters, i.e. https://<fhir-server-url>/Patient/<id> . If they are trying to use the id from the message payload sent to the connector this won’t work. The id URL is an internal id generated by the FHIR server. The provided id we use to generate an identifier on the Patient according to the FHIR spec and needs to be queried (Patient/?identifier=myid)
• The data isn’t being successfully imported. They should be able to use the metrics section in the portal under monitoring. Once there, they should have an IoT Connector tab. Here we have metrics for number of incoming messages, number of normalized messages, and number of message groups. Basically these need to have a non-zero value for each stage. If any is zero for the time period the customer sent the message then the corresponding stage isn’t properly configured. There is also a total error count. They should be able to use this to identify the issue if there are indeed errors.
The most likely cause is the patient or device id isn’t properly configured in the device mapping template (or missing from the payload). Their search on the root Patient resource in combination with the last question leads me to suspect their mapping template or device payload doesn’t have a patient id to extract leading to errors on the FHIR conversion side. Our github page has a list of the different types of issues one can encounter, https://github.com/microsoft/iomt-fhir/blob/master/docs/Debugging.md. The one I suspect is happening at a minimum (though there could be multiple) is PatientIdentityNotDefinedException.
Please check and get back to us with questions.
Thanks
Navtej S