Use better update strategy for konnectivty daemonset

Currently the konnectivity agent has the following update strategy:

```
updateStrategy:
rollingUpdate:
maxUnavailable: 1
maxSurge: 0

```

We (IBM) suggest to update it to the following:
```
updateStrategy:
rollingUpdate:
maxUnavailable: 10%
type: RollingUpdate
```

In a big cluster, it would speed up the konnectivity-agent update. As the agents are independent, this would not hurt the service.