Failure to resolve bad Synced status after it's fixed

We created a Secret `backup-db-simple-crypt-key`, and then a corresponding SealedSecret. The SealedSecret was correctly annotated with:

```yaml
status:
conditions:
- lastTransitionTime: "2021-06-11T21:39:06Z"
lastUpdateTime: "2021-06-11T21:39:06Z"
message: 'failed update: Resource "backup-db-simple-crypt-key" already exists
and is not managed by SealedSecret'
status: "False"
type: Synced
observedGeneration: 2
```

Then we deleted the blocking Secret and restarted sealed-secrets-controller. It correctly created a new Secret from the SealedSecret, but it didn't remove the status. As a result, our health checks (via ArgoCD) are failing even though everything is actually fine.

We're running sealed-secrets-controller with `args: [ "--update-status" ]`.

Reproduce the scenario with a SealedSecret, an existing unmanaged Secret, deletion of that Secret, and the controller running with --update-status. Trace the reconciliation and status-update path after the Secret is recreated; done means the Synced condition no longer reports the old failure and ArgoCD health checks pass.

Written by the indexing model from the issue text.