Dataverse healthcare APIs: Use Healthcare data pipeline template to deploy Azure Logic Apps

This article provides a step-by-step guide for using a template to deploy a group of Azure Logic Apps that ingest FHIR data into Dataverse healthcare APIs, Azure Health Data Services, or both. This solution works as an enterprise-ready Logic App flow that serves as a relay between Azure Health Data Services and Dataverse healthcare APIs. The flow also manages retry logic and exception handling. It relies on an Azure Blob Storage trigger rather than the HTTP trigger used in the manual configuration.

This workflow is available for deployment as an Azure Resource Manager (ARM) template titled Healthcare data pipeline template. You can deploy the template from Microsoft Cloud Solution Center. This offering is a more robust and supported solution provided by Microsoft Cloud for Healthcare. You need to set up some basic manual configuration after deploying the template.

The logic app services rely on an Azure Blob Storage trigger to trigger asynchronous processing of the bundles posted to a configurable storage location. This option handles heavier loads for enterprise use-cases and includes extra exception handling steps. However, you should conduct thorough testing with your expected daily loads.

After deployment, you can extend the logic apps to suit your particular system needs.

The configuration includes the following steps:

• Prerequisites
• Design
• Post deployment steps
• Handle errors
• Set up retries
• Secure Logic Apps

Prerequisites

Ensure that your environment meets the following requirements before you deploy the template:

• An Azure account and subscription. If you don't have a subscription, sign up for a free Azure account before you begin.
• An Azure resource group configured with appropriate permissions to create new resources, or a Contributor role to create new resource groups.
• Access within the resource group to create resources and assign Azure roles.
• Adherence to security guidelines outlined by Azure administrators and organization policy.

Design

The following diagram illustrates the design of the pipeline deployed through the template:

The ARM template deploys several modularized Logic Apps. It includes the following three logic app services:

Template parameters

Resources deployed

The template deploys the following resources in your environment:

Output

Based on whether the run terminates with a success or an error, a blob with the name originalblobname_response.json is created in the bundlesarchive or bundleserror folder with the following schema:

{
  "dataverseResponse": "<The response from the Dataverse healthcare API post the call.>",
  "fhirServerResponse": "<The response from the FHIR server call if the "Post to FHIR server" parameter value was set to True.>",
  "statusMessage": "<Summary of the responses. In case of a failure, the message provides details about how many resources failed to post to the FHIR server and to Dataverse.>",
  "statusCode": "<Code value associated with the issue encountered.>"
}

Depending on which logic app triggered the error, the JSON error contains either the dataverseResponse or fhirServerResponse node. For example, if you encounter an error with the lasendbundletofhir_UniqueValue logic app, the JSON response only contains the fhirServerResponse node and value.

Post deployment steps

The following section contains the steps that you need to follow after deploying the template.

Grant access to the FHIR server

Accessing the FHIR server from the logic app requires the FHIR Data Contributor role assignment, which allows posting new data to the service. Add this Azure role assignment to the managed identity used by the logic app.

1. Go to the FHIR server instance, select Access Control (IAM), and then select Add role assignment.

   In the Role tab, select the FHIR Data Contributor role.

2. Select Members, select Managed identity, and then select + Select members.

3. Add the managed identity created with the ARM template deployment. The newly deployed managed identity should be named mi_UniqueValue.

4. The assignment might take a few minutes to reflect on the managed identity. Select Azure role assignments to view the role assignment on the managed identity.

Grant access to Dataverse healthcare APIs

The same managed identity is used in the logic app to access the Dataverse healthcare APIs by connecting it to an application user in the target Dataverse instance. For more information about application users, go to Manage application users in the Power Platform admin center.

1. You need the Azure Client ID for the managed identity to configure the application user. To retrieve the Client ID, open the managed identity created with the ARM template deployment and copy the Client ID value from the Overview area.

2. In the Power Platform admin center, open the target Microsoft Cloud for Healthcare environment. In the Access section, select S2S Apps, and then select New app user.

3. In the Create a new app user pane, select the appropriate Business unit, and then select Add an app.

4. In the Add an app from Microsoft Entra ID pane, search for the Client ID you copied from the managed identity.

   Select the managed identity from the list, select Add, and then edit the security roles.

5. Select the Sync admin for FHIR App Reg User role, and then select Save.

6. Select Create to create the new application user.

After you complete the configuration, you can test the logic app workflow by posting a sample bundle to the sa_UniqueValue container for processing. Depending on your solution requirement, you can also modify any of these logic apps for more processing.

Handle errors

If a logic app run results in an error, a file with the name originalblobname_response.json is created in the bundleserror container in the storage account. You can parse through this file to identify the root cause of the error, fix it, and resubmit the bundle with the failed resources.

Bundle type: Batch

The FHIR server and the Dataverse healthcare APIs process a bundle of the type batch as a group of independent actions. As a result, the responses indicate the success and failure of each resource independently.

As per the FHIR specification, any resource that fails results in an OperationOutcome with severity value set to error, while the Dataverse healthcare API sets the msind_requeststatus to 935000002. For more information about the request status types, go to Types of request status.

The logic app workflow parses through both the responses from the FHIR server and the Dataverse healthcare APIs. It terminates the flow as Failed if there's any resource that resulted in an error.

Set up retries

After you identify and fix the error, you can place the bundle back in the bundles container for reprocessing.

Retry on throttle: FHIR server

The HTTP action in the logic app workflow that posts to the FHIR server uses the built-in HTTP action retry policies. The default value is an exponential interval policy set to retry four times. You can edit the retry policy.

1. Select the ellipsis in the upper-right corner of the action card, and then select Settings.

2. Under Retry Policy, change the value of the field Type.

   

Retry on throttle: Dataverse healthcare APIs

The Service protection API limits affect the Dataverse healthcare APIs. If a request to Dataverse healthcare APIs gets throttled, the logic app workflow retries three times (by default) at the Retry-After interval specified by the API in the response header. You can edit both the retry count and the interval.

1. To change the retry count, edit the Count value in the Loop until action.

   

2. To change the interval, edit the Count value in the Delay action.

   

Secure Logic Apps

After you finish setting up and testing the logic app, you can lock down the tracing by securing the inputs and outputs actions. To learn more, go to Secure Logic App.

Related information

• Overview of Dataverse healthcare APIs
• Configure Dataverse healthcare APIs
• Dataverse healthcare APIs: Configure Azure Logic App with an HTTP trigger
• Use the Dataverse healthcare APIs