How to config health check of haproxy pod in a kubernetes cluster?

We deploy a Haproxy Pod in kubernetes cluster, and we config health check below:

"livenessProbe": {
   "exec": {
      "command": [
          "test `ps -eo stat,command | grep -v grep | grep haproxy | awk '{print $1}' | grep ^T | wc -l` == 0"
   "initialDelaySeconds": 15,
   "periodSeconds": 3,
   "failureThreshold": 3,
   "timeoutSeconds": 3

In this way, the exec command rely on docker service and does not really check the status of haproxy service. As I know, haproxy does not open an internal port for health check in such k8s pod type.