Bulk exports all resources from the FHIR store to the specified destination. Implements the FHIR implementation guide system level $export. The following headers must be set in the request: * Accept: specifies the format of the OperationOutcome response. Only application/fhir+json is supported. * Prefer: specifies whether the response is immediate or asynchronous. Must be to respond-async because only asynchronous responses are supported. Specify the destination for the server to write result files by setting the Cloud Storage location bulk_export_gcs_destination on the FHIR store. URI of an existing Cloud Storage directory where the server writes result files, in the format gs://{bucket-id}/{path/to/destination/dir}. If there is no trailing slash, the service appends one when composing the object path. The user is responsible for creating the Cloud Storage bucket referenced. Supports the following query parameters: * _type: string of comma-delimited FHIR resource types. If provided, only the resources of the specified type(s) are exported. * _since: if provided, only the resources that are updated after the specified time are exported. * _outputFormat: optional, specify ndjson to export data in NDJSON format. Exported file names use the format: {export_id}_{resource_type}.ndjson. On success, the Content-Location header of the response is set to a URL that the user can use to query the status of the export. The URL is in the format: projects/{project_id}/locations/{location_id}/datasets/{dataset_id}/fhirStores/{fhir_store_id}/operations/{export_id}. See get-fhir-operation-status for more information. Errors generated by the FHIR store contain a JSON-encoded OperationOutcome resource describing the reason for the error.