Remove waiting for pods after the rollout completes.

The wait takes a long time, and seems to get confused if cluster autoscaler
adds/removes nodes and shifts pods around between them. This leads to lengthy
tests that eventually fail with a timeout error. Plus, at least in theory, once
the rollout completes, we should have the pods and shouldn't have to wait at
all... though we suspect this may not always be true.

PiperOrigin-RevId: 822736504

Author: p3rf Team

perfkitbenchmarker/linux_benchmarks/kubernetes_scale_benchmark.py

@@ -204,13 +204,6 @@ def ScaleUpPods(
     cluster.WaitForRollout(rollout_name, timeout=max_wait_time)
 
     all_new_pods = set(cluster.GetPodNames()) - initial_pods
-    cluster.WaitForResource(
-        'pod',
-        condition_name='Ready',
-        timeout=max_wait_time,
-        wait_for_all=True,
-        namespace='default',
-    )
     end_polling_time = time.monotonic()
     logging.info(
         'In %d seconds, found all %s new pods',