It seems we can't get expected "correlation_id" via DjangoRequestReader class.

def get_http_header(self, request, header_name, default=None):
  if request is None:
      return default
  
  if header_name in request.META:
      return request.META.get(header_name)
  if header_name.upper() in request.META:
      return request.META.get(header_name.upper())
  
  return default

Since in Django, header name will be prefixed with "HTTP_" and replace "-" to "_" in request.META.

Is there any special consideration here?

I am glad to create a PR if it needs fix here.