Add node tolerations to kubernetes integration tests

What does this MR do?

I've tainted the worker node pool that was used to schedule jobs and integration tests on so that system pods don't get scheduled on it. If system pods get scheduled on a node the cluster autoscaler wouldn't downscale that node - ever.

A consequence of that is that the pods that we want scheduled on these nodes also need to have tolerations set.

---

As for the updated node selector:

I've updated the selectors on all node pools and this is the new format. The old one is still present but this updates it to keep up.

---

This is a screenshot of the tests pods being scheduled onto the worker pool node:

Why was this MR needed?

What's the best way to test this MR?

What are the relevant issue numbers?