Restarting a Single Container in a Multi-Container Kubernetes Pod

2 min readDec 10, 2023

Kubernetes is the go-to platform for managing containerized applications, allowing you to pack multiple containers into a single “pod” for collaborative execution. However, restarting a single container within a multi-container pod can be tricky, as Kubernetes primarily manages pods as a whole.

The Challenge

Imagine you have a pod with several containers, like Airflow with its worker and git-sync containers. If the git-sync container encounters issues, restarting just that container would be ideal. Unfortunately, Kubernetes doesn’t provide a built-in method for this type of granular control.

Why Restart a Single Container?

Here are some situations where restarting a single container within a multi-container pod is beneficial:

  • Troubleshooting: When a specific container encounters errors, restarting it can resolve the issue without affecting other containers in the pod.
  • Resource Management: Restarting a container that consumes excessive resources can free up resources for other containers in the pod.
  • Scaling Specific Services: In complex microservices architectures, restarting individual services (represented by containers) may be necessary for maintenance or updates.


While a direct command doesn’t exist, you can leverage the power of kubectl and shell scripting to achieve your goal. Here’s the magic command:

# kubectl -n <namespace> exec -it <pod name> -c <container_name> -- /bin/sh -c "kill 1"

Breaking it down:

  • kubectl: Kubernetes command-line tool.
  • -n <namespace>: Replace with your pod’s namespace.
  • exec -it <pod name>: Executes a command in a running container, with “-it” enabling an interactive shell.
  • -c <container_name>: Specifies the container you want to restart.
  • /bin/sh -c "kill 1": Executes a shell command within the container. “kill 1” sends a signal to the primary process (PID 1), effectively restarting it.

Before Restart:

Careful of these considerations:

  • Graceful Termination: Ideally, containers should handle termination gracefully. Sending SIGTERM allows cleanup before shutdown, but not all containers are designed for it.
  • Container Dependencies: Restarting one container might impact others if they share resources or communicate heavily.
  • Monitoring: Implement robust monitoring to detect issues early and automate recovery.
  • Kubernetes Version: Command behavior’s might vary across versions, so consult the relevant documentation.


While native support for single-container restarts is absent, a combination of kubectl and shell scripting can help you achieve this. Remember to consider the implications of container restarts in your specific environment to maintain application stability.

If you like our posts, please share with your friends.

You can subscribe us on

Follow us on Twitter & Instagram




Discuss about #Linux, #DevOps, #Docker, #kubernetes, #HowTo’s, #cloud & IT technologies like #argocd #crossplane #azure