Lower the CPU request for the Affinity Assistant

1/10th of a core may be too much to request. This may affect how much
TaskRun pods can request, especially on single-core Nodes.

This commit lower the CPU request to 1/20th of a core.

/kind misc

pkg/reconciler/pipelinerun/affinity_assistant.go

@@ -140,11 +140,11 @@ func affinityAssistantStatefulSet(name string, pr *v1beta1.PipelineRun, claimNam
 		// Affinity Assistant pod is a placeholder; request minimal resources
 		Resources: corev1.ResourceRequirements{
 			Limits: corev1.ResourceList{
-				"cpu":    resource.MustParse("100m"),
+				"cpu":    resource.MustParse("50m"),
 				"memory": resource.MustParse("100Mi"),
 			},
 			Requests: corev1.ResourceList{
-				"cpu":    resource.MustParse("100m"),
+				"cpu":    resource.MustParse("50m"),
 				"memory": resource.MustParse("100Mi"),
 			},
 		},