Single node reserve 3GiB of system memory to avoid alerts

ci-operator/step-registry/openshift/e2e/aws/single-node/openshift-e2e-aws-single-node-workflow.yaml

@@ -4,6 +4,7 @@ workflow:
     pre:
     - chain: ipi-conf-aws
     - ref: single-node-conf-aws
+    - ref: single-node-conf-e2e
     - chain: ipi-install
     test:
     - ref: openshift-e2e-test

ci-operator/step-registry/openshift/e2e/gcp/single-node/openshift-e2e-gcp-single-node-workflow.yaml

@@ -4,6 +4,7 @@ workflow:
     pre:
     - chain: ipi-conf-gcp
     - ref: single-node-conf-gcp
+    - ref: single-node-conf-e2e
     - chain: ipi-install
     test:
     - ref: openshift-e2e-test

ci-operator/step-registry/single-node/conf/e2e/OWNERS

@@ -0,0 +1,7 @@
+approvers:
+- romfreiman
+- eranco74
+- tsorya
+- osherdp
+- yuvigold
+- omertuc

ci-operator/step-registry/single-node/conf/e2e/single-node-conf-e2e-commands.sh

@@ -0,0 +1,21 @@
+#!/bin/bash
+
+set -o errexit
+set -o nounset
+set -o pipefail
+
+# Reserve more system memory per node than a typical multi-node cluster
+# to facilitate E2E tests on a single node.
+cat > "${SHARED_DIR}/manifest_single-node-reserve-sys-mem-kubeletconfig.yml" << EOF
+apiVersion: machineconfiguration.openshift.io/v1
+kind: KubeletConfig
+metadata:
+  name: single-node-reserve-sys-mem
+spec:
+  machineConfigPoolSelector:
+    matchLabels:
+      pools.operator.machineconfiguration.openshift.io/master: ""
+  kubeletConfig: 
+    systemReserved:
+      memory: 3Gi
+EOF

ci-operator/step-registry/single-node/conf/e2e/single-node-conf-e2e-ref.metadata.json

@@ -0,0 +1,13 @@
+{
+	"path": "single-node/conf/e2e/single-node-conf-e2e-ref.yaml",
+	"owners": {
+		"approvers": [
+			"romfreiman",
+			"eranco74",
+			"tsorya",
+			"osherdp",
+			"yuvigold",
+			"omertuc"
+		]
+	}
+}

ci-operator/step-registry/single-node/conf/e2e/single-node-conf-e2e-ref.yaml

@@ -0,0 +1,10 @@
+ref:
+  as: single-node-conf-e2e
+  from: installer
+  commands: single-node-conf-e2e-commands.sh
+  resources:
+    requests:
+      cpu: 10m
+      memory: 100Mi
+  documentation: |-
+      Configurations required to run E2E tests on a single node cluster.