smallcase/argoworkflow-pod-reaper

Deletes completed pods that are owned by ArgoWorkflow.

argoworkflow-pod-reaper

Deletes completed pods that are owned by ArgoWorkflow.

Usage:

go test ./... -cover        
ok      github.com/smallcase/workflow-reaper    0.008s  coverage: 48.6% of statements

go build -o app 

./app --in-cluster=false --delete-failed-after=1 --delete-successful-after=1 --namespaces="ns1,ns2"

Example command output:

2021/08/24 20:43:06 Deleting pods that failed 1 days ago
2021/08/24 20:43:06 Deleting pods that succeeded 1 days ago
2021/08/24 20:43:06 Would delete pod static-tester in namespace ns1
2021/08/24 20:43:06 Would delete pod static-tester in namespace ns2

Container image build:

chmod +x build.sh
./build.sh

Kubernetes job example:

apiVersion: batch/v1
kind: Job
metadata:
  generateName: example-job
spec:
  template:
    spec:
      containers:
      - name: example
        image: $MY_IMAGE:$TAG
      restartPolicy: Never
      serviceAccountName: example
  backoffLimit: 0

ArgoWorkflow example:

apiVersion: argoproj.io/v1alpha1
kind: Workflow                  
metadata:
  generateName: hello-world-    
spec:
  serviceAccountName: example
  entrypoint: example          
  templates:
  - name: example             
    container:
      image: $MY_IMAGE:$TAG

Required RBAC:

apiVersion: v1
kind: ServiceAccount
metadata:
  name: example
  namespace: default
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: example
rules:
- apiGroups:
  - ""
  resources:
  - pods
  verbs:
  - get
  - list
  - patch
  - watch
  - delete
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: example
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: example
subjects:
- kind: ServiceAccount
  name: example
  namespace: default

---

flag	default	type	usage
--in-cluster	true	bool	use in cluster config, or ~/.kube/config
--delete-failed-after	10	int	delete failed pods after x days
--delete-successful-after	5	int	delete successful pods after x days
--namespaces	default	[]string	namespaces to watch
--dry-run	true	bool	dry run pod deletion