Portal:Toolforge/Admin/Runbooks/ToolforgeKubernetesNodeNotReady

The procedures in this runbook require admin permissions to complete.

The ToolforgeKubernetesCapacity alert fires when a Toolforge Kubernetes node is marked as not ready. A paging alert also fires when at least 5 nodes are marked as not ready.

Debugging

On a bastion run as your own user:

$ kubectl sudo get node
$ kubectl sudo describe node <node>

Support contacts

Old incidents

This article is issued from Wikimedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.