Create README.md (#1035)

* Create README.md * Update scripts/cluster_cleanup/README.md Co-authored-by: Nikos Kontakis <wirednkod@gmail.com> * Update scripts/cluster_cleanup/README.md Co-authored-by: Nikos Kontakis <wirednkod@gmail.com> * Update scripts/cluster_cleanup/README.md Co-authored-by: Nikos Kontakis <wirednkod@gmail.com> * Update scripts/cluster_cleanup/README.md Co-authored-by: Nikos Kontakis <wirednkod@gmail.com> --------- Co-authored-by: Nikos Kontakis <wirednkod@gmail.com>
paritytech · May 17, 2023 · b0bef5e · b0bef5e
1 parent b18c7d5
commit b0bef5e
Showing 1 changed file with 42 additions and 0 deletions.
diff --git a/scripts/cluster_cleanup/README.md b/scripts/cluster_cleanup/README.md
@@ -0,0 +1,42 @@
+### Zombie Namespace Cleanup Script
+
+This Python script is designed to identify and delete "zombie-" prefixed Kubernetes namespaces and PodMonitors that are older than 12 hours of age. It uses the Kubernetes Python client library to connect to the API server and query the list of namespaces and podmonitors and then proceeds with deleting them.
+
+### Usage
+
+- Ensure that the Kubernetes Python client library is installed (`pip install -r requirements.txt`);
+- Update the `prefix` and `time_delta` variables in the script to match the desired settings;
+- Run the script with python `cluster_cleanup.py`;
+
+### CronJob
+
+To automate the zombie namespace cleanup process, the following Kubernetes CronJob can be used:
+
+```
+apiVersion: batch/v1
+kind: CronJob
+metadata:
+  name: cluster-cleanup
+  namespace: cluster-cleanup
+spec:
+  schedule: "0 * * * *"
+  jobTemplate:
+    spec:
+      template:
+        spec:
+          serviceAccountName: cluster-cleanup
+          containers:
+          - name: cluster-cleanup
+            image: emamihe/cluster-cleanup
+          restartPolicy: OnFailure
+```
+
+This CronJob will run the emamihe/cluster-cleanup image with the cluster-cleanup service account every hour in the cluster-cleanup namespace. 
+
+To deploy the CronJob, save the manifest to a file (e.g. cronjob.yaml) and run the following command:
+
+```
+kubectl apply -f cronjob.yaml
+```
+
+This will create the CronJob in the cluster and start running it on the specified schedule.