RDC Prometheus plugin return errors when use the --rdc_gpu_indexes

When above option is used, the plugin returns errors:
  result = rdc.rdc_group_gpu_add(rdc_handle, gpu_group_id, gpu)
  ctypes.ArgumentError: argument 3: <type 'exceptions.TypeError'>: wrong type

The rdc_prometheus.py is changed to convert string to integer.
The RdcUtil.py is also changed to raise Exception properly.

Change-Id: I9535091ff1fc8882cccd32e5f2810da5241768c3


[ROCm/rdc commit: 7ca7a571a7]

projects/rdc/python_binding/rdc_prometheus.py

@@ -88,6 +88,9 @@ if __name__ == '__main__':
     rdc_ip_port = args.rdc_ip_port
     if args.rdc_embedded:
         rdc_ip_port = None
+    if args.rdc_gpu_indexes != None:
+        for i in range(0, len(args.rdc_gpu_indexes)):
+            args.rdc_gpu_indexes[i] = int(args.rdc_gpu_indexes[i])
 
     reader = PrometheusReader(rdc_ip_port, field_ids, args.rdc_update_freq*1000000,
             args.rdc_max_keep_age, args.rdc_max_keep_samples,