test_clusterip_service_endpoint fails on connection error in k8s 1.24 on ARM

In testrun https://solutions.qa.canonical.com/testruns/testRun/5a82c72a-b16d-449a-82ac-665c660aa4c2, with FCE output https://oil-jenkins.canonical.com/job/fce_build/2153//console we see test_clusterip_service_endpoint fail:

```
=================================== FAILURES ===================================
_______________________ test_clusterip_service_endpoint ________________________
Traceback (most recent call last):
  File "/home/ubuntu/k8s-validation/jobs/integration/test_service_endpoints.py", line 124, in test_clusterip_service_endpoint
    raise e
  File "/home/ubuntu/k8s-validation/jobs/integration/test_service_endpoints.py", line 121, in test_clusterip_service_endpoint
    assert "Hello Kubernetes!" in action.results.get("Stdout", "")
AssertionError: assert 'Hello Kubernetes!' in ''
 + where '' = <built-in method get of dict object at 0x7fd2bfb0f440>('Stdout', '')
 + where <built-in method get of dict object at 0x7fd2bfb0f440> = {'Code': '7', 'Stderr': '* Trying 10.152.183.132:80...\n* TCP_NODELAY set\n % Total % Received % Xferd Average Speed Time Time Time Current\n Dload Upload Total Spent Left Speed\n\r 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0* connect to 10.152.183.132 port 80 failed: Connection refused\n* Failed to connect to 10.152.183.132 port 80: Connection refused\n\r 0 0 0 0 0 0 0 0 --:--:-- 0:00:01 --:--:-- 0\n* Closing connection 0\ncurl: (7) Failed to connect to 10.152.183.132 port 80: Connection refused\n'}.get
 + where {'Code': '7', 'Stderr': '* Trying 10.152.183.132:80...\n* TCP_NODELAY set\n % Total % Received % Xferd Average Speed Time Time Time Current\n Dload Upload Total Spent Left Speed\n\r 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0* connect to 10.152.183.132 port 80 failed: Connection refused\n* Failed to connect to 10.152.183.132 port 80: Connection refused\n\r 0 0 0 0 0 0 0 0 --:--:-- 0:00:01 --:--:-- 0\n* Closing connection 0\ncurl: (7) Failed to connect to 10.152.183.132 port 80: Connection refused\n'} = <Action entity_id="26">.results
------------------------------ Captured log call -------------------------------

```

The bundle we are using is https://oil-jenkins.canonical.com/artifacts/5a82c72a-b16d-449a-82ac-665c660aa4c2/generated/generated/kubernetes-maas/bundle.yaml and these are ARM64 machines. We do not see this problem on AMD64 machines, nor was this happening on ARM64 k8s 1.23.

In the logs we see that this IP belongs to the service `default/hello-world`:

```
3/lxd/1/var/log/syslog-May 18 21:48:19 juju-615a59-3-lxd-1 kube-apiserver.daemon[371087]: I0518 21:48:19.883484 371087 httplog.go:131] "HTTP" verb="LIST" URI="/api/v1/namespaces/default/resourcequotas" latency="2.927089ms" userAgent="kube-apiserver/v1.24.0 (linux/arm64) kubernetes/4ce5a89" audit-ID="cc8735d9-3e98-4635-9802-79c8355b75b6" srcIP="[::1]:35452" apf_pl="exempt" apf_fs="exempt" apf_execution_time="2.691968ms" resp=200
3/lxd/1/var/log/syslog:May 18 21:48:19 juju-615a59-3-lxd-1 kube-apiserver.daemon[371087]: I0518 21:48:19.887041 371087 alloc.go:327] "allocated clusterIPs" service="default/hello-world" clusterIPs=map[IPv4:10.152.183.132]
3/lxd/1/var/log/syslog-May 18 21:48:19 juju-615a59-3-lxd-1 kube-apiserver.daemon[371087]: I0518 21:48:19.887521 371087 httplog.go:131] "HTTP" verb="POST" URI="/api/v1/namespaces/default/services?fieldManager=kubectl-expose" latency="16.499693ms" userAgent="kubectl/v1.24.0 (linux/amd64) kubernetes/4ce5a89" audit-ID="2914d3d0-d685-421b-8d8b-58010b6b1d25" srcIP="10.246.200.115:41676" apf_pl="exempt" apf_fs="exempt" apf_execution_time="15.977931ms" resp=201
```

But there is no further indication why it can't connect.

Logs can be found here:
https://oil-jenkins.canonical.com/artifacts/5a82c72a-b16d-449a-82ac-665c660aa4c2/index.html