[designate] Remove RabbitMQ connection check from API health-probe

Unfortunately, neither RabbitMQ connection is established
nor heartbeats are working before any request arrives
to the wsgi apache process.

So for this check to work some of the following is required:
- Constant stream of incoming requests triggering the connection to Rabbit
- API should run as a native eventlet application (designate-api)
- API should run as uwsgi application with pthreads enabled for heartbeats

Neither of these options is good for us, so disabling this part of health
probe for API service for now.

openstack/designate/Chart.yaml

@@ -1,7 +1,7 @@
 apiVersion: v2
 description: A Helm chart for Kubernetes to deploy Openstack Designate  (DNSaaS)
 name: designate
-version: 0.4.0
+version: 0.4.1
 appVersion: "xena"
 dependencies:
   - condition: percona_cluster.enabled

openstack/designate/bin/health-probe.py

@@ -19,8 +19,7 @@
 Uses oslo's ping method that is designed just for such simple purpose.
 
 For API service script:
-  a. checks if there are any established RabbitMQ connections in a pod.
-  b. checks if the RabbitMQ and DB are reachable over TCP.
+  a. checks if the RabbitMQ and DB are reachable over TCP.
 These two checks succeed or timeout within 5 seconds.
 
 For other RPC services script:
@@ -138,21 +137,6 @@ def tcp_socket_status(process: Optional[str], ports: Set[int]) -> int:
     return 0
 
 
-def tcp_connection_established(ports: Set[int]) -> int:
-    """Check the presence of established TCP connections to remote ports"""
-    scon = psutil.net_connections(kind="tcp")
-    for con in scon:
-        try:
-            rport = con.raddr[1]
-            status = con.status
-        except IndexError:
-            # Skip if raddr is empty tuple: connection is not ESTABLISED or connection is LISTEN
-            continue
-        if rport in ports and status == tcp_established:
-            return 1
-    return 0
-
-
 def check_tcp_socket(
     service: str,
     rabbits: Set[Tuple[str, int]],
@@ -189,18 +173,6 @@ def check_tcp_socket(
                     sys.exit(1)
 
 
-def check_rabbitmq_tcp_socket(rabbits: Set[Tuple[str, int]]) -> None:
-    """Check RabbitMQ TCP socket in ESTABLISHED state is present"""
-    r_ports = {port for _, port in rabbits}
-    # Check that there are any established RabbitMQ connections
-    # Apache2 wsgi RabbitMQ connections doesn't belong to any process in pods netstat
-    if r_ports and tcp_connection_established(r_ports) == 0:
-        sys.stderr.write(f"RabbitMQ socket not established")
-        # Do not kill the pod if RabbitMQ is not reachable/down
-        if not cfg.CONF.liveness_probe:
-            sys.exit(1)
-
-
 def check_tcp_connectivity(
     rabbits: Set[Tuple[str, int]],
     databases: Set[Tuple[str, int]],
@@ -292,7 +264,6 @@ def test_liveness() -> None:
         check_service_status(transport)
 
     if service in api_services:
-        check_rabbitmq_tcp_socket(rabbits)
         check_tcp_connectivity(rabbits, databases)