cluster-api-provider-azure: NMI in a crashloop

/kind bug

What steps did you take and what happened: Created a CAPZ cluster and noticed that the NMI pods are in a constant crashloop.

What did you expect to happen: NMI should not crashloop.

Anything else you would like to add: Log from a pod:

I0720 16:34:18.943358       1 main.go:61] starting nmi process. Version: v1.6.3. Build date: 2020-09-15-21:07.
I0720 16:34:19.046225       1 crd.go:446] CRD lite informers started
I0720 16:34:19.146391       1 main.go:88] running NMI in namespaced mode: true
I0720 16:34:19.146446       1 nmi.go:53] initializing in managed mode
I0720 16:34:19.146454       1 probes.go:41] initialized health probe on port 8085
I0720 16:34:19.146460       1 probes.go:44] started health probe
F0720 16:34:19.146873       1 probes.go:27] http listen and serve error: listen tcp :8085: bind: address already in use

Environment:

• cluster-api-provider-azure version: 0.4.15
• Kubernetes version: (use kubectl version):

Client Version: version.Info{Major:"1", Minor:"21", GitVersion:"v1.21.0", GitCommit:"cb303e613a121a29364f75cc67d3d580833a7479", GitTreeState:"clean", BuildDate:"2021-04-08T21:10:45Z", GoVersion:"go1.16.3", Compiler:"gc", Platform:"darwin/arm64"}
Server Version: version.Info{Major:"1", Minor:"20", GitVersion:"v1.20.7", GitCommit:"132a687512d7fb058d0f5890f07d4121b3f0a2e2", GitTreeState:"clean", BuildDate:"2021-05-12T12:32:49Z", GoVersion:"go1.15.12", Compiler:"gc", Platform:"linux/amd64"}

• OS (e.g. from /etc/os-release): default