Manually stop or start containers in Azure Container Instances

The restart policy setting of a container group determines how container instances start or stop by default. You can override the default setting by manually stopping or starting a container group.

Stop

Manually stop a running container group - for example, by using the az container stop command or Azure portal. For certain container workloads, you might want to stop a long-running container group after a defined period to save on costs.

When a container group enters the Stopped state, it terminates and recycles all the containers in the group. It does not preserve container state.

When the containers are recycled, the resources are deallocated and billing stops for the container group.

The stop action has no effect if the container group already terminated (is in either a Succeeded or Failed state). For example, a container group with run-once container tasks that ran successfully terminates in the Succeeded state. Attempts to stop the group in that state do not change the state.

Start

When a container group is stopped - either because the containers terminated on their own or you manually stopped the group - you can start the containers. For example, use the az container start command or Azure portal to manually start the containers in the group. If the container image for any container is updated, a new image is pulled.

Starting a container group begins a new deployment with the same container configuration. This action can help you quickly reuse a known container group configuration that works as you expect. You don't have to create a new container group to run the same workload.

All containers in a container group are started by this action. You can't start a specific container in the group.

After you manually start or restart a container group, the container group runs according to the configured restart policy.

Restart

You can restart a container group while it is running - for example, by using the az container restart command. This action restarts all containers in the container group. If the container image for any container is updated, a new image is pulled.

Restarting a container group is helpful when you want to troubleshoot a deployment problem. For example, if a temporary resource limitation prevents your containers from running successfully, restarting the group might solve the problem.

All containers in a container group are restarted by this action. You can't restart a specific container in the group.

After you manually restart a container group, the container group runs according to the configured restart policy.

Next steps

Learn more about restart policy settings in Azure Container Instances.

In addition to manually stopping and starting a container group with the existing configuration, you can update the settings of a running container group.