Post-Deployment Operations

Checking application logs

To check the logs of a particular service, we can use the built-in mechanisms and using the kubectl tool we can check the logs in the following ways:

Basic log retrieval

Retrieve logs from a specific pod:

kubectl logs $pod-name -n $namespace

Follow mode (real-time logs)

Monitor logs in real-time as they are generated:

kubectl logs -f $pod-name -n $namespace

Using labels:

kubectl logs -l $label-selector -n $namespace
kubectl logs -l app.kubernetes.io/name=som-dev-service-validation -n $namespace
kubectl logs -f -l app.kubernetes.io/name=som-dev-service-validation -n $namespace

Logs within a time interval:

Retrieve logs from the last specified time interval:

kubectl logs $pod-name --since=10m -n $namespace

Restart application

kubectl rollout restart is the preferred method for restarting an application, as it initiates a controlled restart of all pods in a given deployment without interrupting the application. This method makes it easy to update the application, allowing it to remain available by gradually replacing old pods with new ones.

kubectl delete pod command removes the selected pods manually, allowing you to restart the pod immediately without running a full rollout process. When a pod is deleted, the ReplicaSet controller will automatically create a new pod to reach the defined number of replicas.

Comparison of Methods: kubectl rollout restart vs kubectl delete pod