Use context: kubectl config use-context k8s-c2-AC

Write a command into /opt/course/15/cluster_events.sh which shows the latest events in the whole cluster, ordered by time (metadata.creationTimestamp). Use kubectl for it.

Now delete the kube-proxy Pod running on node cluster2-node1 and write the events this caused into /opt/course/15/pod_kill.log.

Finally kill the containerd container of the kube-proxy Pod on node cluster2-node1 and write the events into /opt/course/15/container_kill.log.

Do you notice differences in the events both actions caused?

Understanding and Managing Kubernetes Events: Pod and Container Termination

In Kubernetes, events provide critical insights into the state of your cluster and its resources. When Pods or containers are terminated, Kubernetes generates events that help administrators understand what actions were taken to maintain the desired state. In this guide, we’ll explore how to generate and capture events related to Pod and container termination, and analyze the differences in the events generated.

Step 1: Capturing Cluster Events

To start, we’ll capture all events across the cluster, sorted by their creation time. This will give us a chronological view of what’s happening in the cluster:

This script lists all events in the cluster and sorts them by the time they were created, making it easier to trace the sequence of actions.

Step 2: Deleting a kube-proxy Pod

Next, we’ll delete a specific kube-proxy Pod to observe the events generated by this action. First, identify the Pod running on cluster2-node1:

Once you’ve identified the Pod, delete it:

After deleting the Pod, run the cluster events script to capture the events triggered by this action:

Step 3: Logging the Events

Write the events caused by the Pod deletion into a log file for further analysis:

This log file contains all the events related to the deletion and recreation of the kube-proxy Pod, including the actions taken by the DaemonSet to recreate the Pod.

Step 4: Killing a Container Inside the Pod

To further analyze event generation, we’ll manually kill the main container inside the kube-proxy Pod. First, SSH into the node where the Pod is running:

Next, find the container ID for the kube-proxy container:

Once you have the container ID, remove the container:

You should see that Kubernetes immediately recreates the container to maintain the Pod’s desired state:

Step 5: Logging Container Kill Events

Now, check and log the events caused by killing the container:

These events show that Kubernetes recreated the container within the existing Pod, with fewer actions required compared to deleting the entire Pod.

Analyzing the Differences

Comparing the events generated by deleting a Pod versus killing a container reveals key differences:

• Pod Deletion: Deleting the entire Pod triggers the DaemonSet to recreate the Pod, resulting in more events, such as scheduling, pulling the image, and starting the Pod.
• Container Kill: Killing a container within a Pod causes fewer events since the Pod still exists, and only the container needs to be recreated.

These differences highlight how Kubernetes efficiently manages resources and ensures that the desired state of your applications is maintained.

Conclusion

In this guide, we’ve explored how Kubernetes handles Pod and container termination, and how these actions generate different events. Understanding these events is crucial for troubleshooting and managing your cluster effectively. By capturing and analyzing these events, you can gain deeper insights into how Kubernetes maintains the desired state of your applications.