The Tensorboard Controller watches and manages Tensorboard custom resources that the Tensorboards Web App creates. For each Tensorboard the user creates, the Tensorboard Controller syncs all Kubernetes resources that are needed to run a Tensorboard instance on Kubernetes. Please refer to the following commands to perform a health check or investigation into the Kubeflow Tensorboard Controller.
1. Check the status of the Tensorboard Controller Deployment:
$ kubectl get deployment -n kubeflow tensorboard-controller-controller-manager -o wide
2. Inspect the events on the Tensorboard Controller Deployment:
$ kubectl describe deployment -n kubeflow tensorboard-controller-controller-manager
3. Check the status of the Tensorboard Controller Pod:
$ kubectl get pods -n kubeflow -l app=tensorboard-controller -o wide
4. View Tensorboard Controller logs:
$ kubectl logs -n kubeflow deploy/tensorboard-controller-controller-manager -c manager -f
Comments
0 comments
Article is closed for comments.