sylva-units HelmRelease remediation settings remain after upgrade from 1.1.1

after an upgrade from 1.1.1, I noticed by pure chance that the sylva-units HelmRelease has the following remediation settings:

spec:
...
    install:
      remediation:
        remediateLastFailure: false
        retries: 2
...
    upgrade:
      remediation:
        retries: 3
...

We don't have such settings on a fresh install, and we do not want them (we removed them in !2469 (merged)).

The issue is that the following happens:

• sylva 1.1.1 bootstrap cluster is installed
• sylva-units management-sylva-units creates the sylva-units HelmRelease in the mgmt cluster
• when it does this it inherits from the unit_helmrelease_spec_default of Sylva 1.1.1, which have the remediation things
• when we upgrade to 1.2 with apply.sh, kubectl redefines the HelmRelease but the previous settings are not removed (I believe that this is because they are associated with a different server-side-apply manager)

Those settings are problematic to keep, because they may result in uninstallations of sylva-units HelmRelease !