To understand the lifecycle of a pod, we will look a the following<\/p>\n
- \n
- Pod Phases<\/li>\n
- Pod Conditions<\/li>\n
- Container Status<\/li>\n<\/ol>\n
To understand the concepts better, let’s assume a multi-container pod with the following<\/p>\n
- \n
- An init container to fetch an API secret in the runtime.<\/li>\n
- container-01 (java-api) runs the java application<\/strong><\/li>\n
- container-02 (log-reader) fetches the application logs and sends them to a log forwarder.<\/li>\n<\/ol>\n
We will understand the pod lifecycle by understanding what happens when we deploy the pod with the above requirements.<\/p>\n
Pod Phases<\/h2>\n
When you deploy a pod, it could typically fall under any one of the following phases.<\/p>\n
<\/p>\n
\n\n
\n \nPo Phases<\/th>\n Description<\/th>\n<\/tr>\n<\/thead>\n \n Pending<\/strong><\/td>\n The Pod is created but not yet running.<\/td>\n<\/tr>\n \n Running<\/strong><\/td>\n At least one container is running, or is in the process of starting or restarting.<\/td>\n<\/tr>\n \n Succeeded<\/strong><\/td>\n All containers have been completed successfully.<\/td>\n<\/tr>\n \n Failed<\/strong><\/td>\n At least one container has failed.<\/td>\n<\/tr>\n \n Unknown<\/strong><\/td>\n The Pod status couldn’t be obtained by the API server.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n <\/p>\n
Now let us understand these phases with a real-time example<\/strong> using our application pod. Let’s call our pod java-api-pod<\/strong><\/p>\n
Here is how the pod phases work when you deploy a pod multi-container pod.<\/p>\n
1. Pending Phase<\/h3>\n
When you deploy the api-pod, it will be in the pending phase. Before the Pod moves to the “Running<\/strong>” phase, the init container will run to completion<\/strong> before any other containers start.<\/p>\n
Following are some of the common scenarios where a Pod continues to be in the pending stage.<\/p>\n
- \n
- Lack of and CPU & Memory resources for the pod.<\/li>\n
- If the pod has a volume definition and the volume is not available.<\/li>\n
- If Kubernetes can’t pull the container image.<\/li>\n
- If an init container fails to start<\/li>\n<\/ol>\n
If the init container fails with a non-zero exit code and the restartPolicy<\/strong> is set to
Never<\/code><\/strong>, the pod directly goes to the failed phase.<\/p>\n2. Running Phase<\/h3>\n
After the init container finishes fetching the secret, the Pod moves to the “Running” phase. Both container-01<\/strong> and container-02<\/strong> will start. Because our Java application is an API that needs to be running, the Pod will stay in the “Running<\/strong>” phase as long as the Java API application is up. For some reason, if any of the containers fail to start, the pod goes to the failed phase.<\/p>\n
3. Succeeded Phase:<\/h3>\n
This phase is not applicable for our Java application, because “Succeeded<\/strong>” is for containers that complete their tasks and then exit.<\/p>\n
Our java-api-pod<\/strong> is meant to keep running, so it won’t reach this phase unless you manually stop it.<\/p>\n
The succeeded phase is applicable for pods that are part of Kubernetes Job\/Cronjob Objects.<\/p>\n
4. Failed Phase<\/strong><\/h3>\n
If your init container, Java application container, or the log reader container crashes or exits for some reason, the Pod will move to the “Failed” phase.<\/p>\n
A pod moves to the failed phase in the following scenarios<\/p>\n
- \n
- If the init container or main container exits with a non-zero exit code and has restartPolicy<\/strong> set to Never<\/strong>.<\/li>\n
- If a node fails or a pod is evicted from a node and can’t be moved to another node, it moves to a failed state.<\/li>\n
- If a pod has activeDeadlineSeconds<\/strong> (Typically on Jobs & Cronjobs<\/a>) field enabled and it exceeds the time limits, the pod gets terminated and marked as Failed.<\/li>\n
- If you manually delete the pod and it can’t gracefully terminate, the pod moves to a failed state.<\/li>\n<\/ol>\n
5. Unknown Phase<\/h3>\n
This is rare, but if the api-server can’t get the status of your Pod for some reason, it will be marked as “Unknown.”<\/p>\n
Pod Conditions<\/h2>\n
The pod’s phase gives a brief update on the pod’s current status as Pod conditions <\/strong>give you detailed information related to scheduling, readiness, and initialization.<\/p>\n
- If the init container or main container exits with a non-zero exit code and has restartPolicy<\/strong> set to Never<\/strong>.<\/li>\n
- container-02 (log-reader) fetches the application logs and sends them to a log forwarder.<\/li>\n<\/ol>\n
![\"Pod](\"https:\/\/storage.ghost.io\/c\/5f\/2f\/5f2f4d20-2abf-4534-8d40-7aa233aedd43\/content\/images\/2025\/03\/image-8-38.png\")
<\/figure>\n