Upgrading sandbox doesn't update cluster code

[daniel-sanche]

When we install sandbox over top of an existing sandbox, we expect the old sandbox to be updated to use the new code. Instead, the cluster seems to continue to run the old container images. This caused issues with the log4j exploit. Users would update their cluster expected to get the patch, but would have to manually run "kubectl apply" to update the cluster code.

Alternatively, I think we should consider removing support for updating existing sandboxes