@leverege/build-tools 2.34.0 → 2.34.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@leverege/build-tools",
-  "version": "2.34.0",
+  "version": "2.34.1",
   "description": "A collection of build / support tools for Leverege developers",
   "main": "index.js",
   "repository": {

package/src/helm-charts/prometheus/values-local.yaml

@@ -197,12 +197,12 @@ serverFiles:
 
           - alert: KubernetesPodNotHealthy
             expr: kube_pod_status_phase{phase=~"Pending|Unknown|Failed"} == 1
-            for: 10m
+            for: 20m
             labels:
               severity: error
             annotations:
               summary: "Kubernetes Pod not healthy"
-              description: "Pod {{ $labels.pod }} has been in a non-ready state for longer than 10 minutes."
+              description: "Pod {{ $labels.pod }} has been in a non-ready state for longer than 20 minutes."
               dashboard: "Check the pod status using the Kubernetes UI"
 
           - alert: KubernetesPodCrashLooping

package/src/helmup

@@ -317,40 +317,35 @@ function installPreemptibleKiller() {
 }
 
 function installCronZombieKiller() {
-  local BATCH_SYNTAX='batch/v1' # GKE 1.21 and beyond
-  [ ! -z "$PRE_GKE_21" ] && BATCH_SYNTAX='batch/v1beta1'
   cat<<END_OF_ZOMBIE | kubectl apply -f -
 apiVersion: v1
 kind: ServiceAccount
 metadata:
   name: cron-zombie-sa
-  namespace: default
 ---
 apiVersion: rbac.authorization.k8s.io/v1
-kind: Role
+kind: ClusterRole
 metadata:
-  name: pod-accessor-role
-  namespace: default
+  name: cron-zombie-role
 rules:
 - apiGroups: [""]
   resources: ["pods"]
-  verbs: ["get","delete","watch","list"]
+  verbs: ["list", "delete"]
 ---
 apiVersion: rbac.authorization.k8s.io/v1
-kind: RoleBinding
+kind: ClusterRoleBinding
 metadata:
-  name: pod-access
-  namespace: default
+  name: cron-zombie-role-binding
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: cron-zombie-role
 subjects:
 - kind: ServiceAccount
   name: cron-zombie-sa
   namespace: default
-roleRef:
-  kind: Role
-  name: pod-accessor-role
-  apiGroup: ""
 ---
-apiVersion: $BATCH_SYNTAX
+apiVersion: batch/v1
 kind: CronJob
 metadata:
   namespace: default
@@ -375,22 +370,9 @@ spec:
               - "/bin/sh"
             args:
               - "-c"
-              - "echo Zombie Hunting `date` ; kubectl get pods -n default --field-selector='status.phase==Failed' -o name | xargs kubectl delete -n default 2> /dev/null"
+              - "echo 'Hunting Zombies' ; kubectl get pods --all-namespaces --field-selector=status.phase=Failed -o json | kubectl delete -f - 2> /dev/null"
 status: {}
 END_OF_ZOMBIE
-
-  if [[ $? -eq 1 ]];
-  then
-    cat<<OLDER_GKE
-
-  Deployment failed is the GKE version pre 1.21? Try rerunning with PRE_GKE_21 defined:
-
-     $ `color y "PRE_GKE_21=1 helmup zombie-killer"`
-
-
-OLDER_GKE
-  exit 1
-  fi
 }
 
 function installCertManagerSecret() {