Skip to main content

Dashboard

You want to: find your workload and see at a glance whether it is healthy.

The Dashboard shows one namespace of your current cluster (Choosing a cluster and namespace). It refreshes on its own every 10 seconds; Refresh (or r) does it now.

The Dashboard on the Pods tab of namespace shop: five pods with their status (Running, Error, OOMKilled), containers, restarts, controller, CPU, memory and ageThe Dashboard on the Pods tab of namespace shop: five pods with their status (Running, Error, OOMKilled), containers, restarts, controller, CPU, memory and age

Tabs​

Pods, Deployments, DaemonSets, StatefulSets, HPAs, Services, ConfigMaps, Secrets, Ingresses, CronJobs and Jobs; you see the ones you may list. [ and ] move to the previous and next tab (on a Turkish keyboard ğ and ü).

The Deployments, StatefulSets and DaemonSets tabs mark a workload with a red N pods failing badge when one of its pods is crashing:

The Deployments tab: catalog and checkout are Progressing with a red '1 pod failing' badge, storefront is Available with 2/2 replicasThe Deployments tab: catalog and checkout are Progressing with a red '1 pod failing' badge, storefront is Available with 2/2 replicas

List or cards​

The two buttons next to the search switch between a table and cards. The table sorts by any column (click its header). Cards show the same facts with the action buttons on each card:

The Pods tab as cards: each pod with status, start time, restarts, controller, CPU and memory, and Logs / Events / Shell buttonsThe Pods tab as cards: each pod with status, start time, restarts, controller, CPU and memory, and Logs / Events / Shell buttons

Columns​

In the table, the ⋮ at the right end of the header row chooses which columns show, per tab. Name always stays; Reset to default shows all of them again. The choice is kept in this browser only.

The column menu open on the Pods tab: checkboxes for Status, Containers, Restarts, Controlled By, CPU, Memory and Age, and Reset to defaultThe column menu open on the Pods tab: checkboxes for Status, Containers, Restarts, Controlled By, CPU, Memory and Age, and Reset to default

Search and label filters​

Press / (or click the search box) and type. Every word you type must match, in any order, ignoring case; part of a word is enough. A word matches the name or the text of any column, hidden columns included, so the table and the cards find the same rows:

TabAlso matched
Podsstatus, container names, Controlled By (kind and name)
Deployments, StatefulSets, DaemonSetsstatus, replicas (2/3)
Servicestype, cluster IP, external IPs, ports (80, 8080/TCP, target port, node port, port name)
ConfigMaps, Secretskey names, secret type
Ingresseshosts and paths
CronJobs, Jobsschedule, status
HPAstarget (Deployment/web), min / max

Ages, restarts, CPU, memory and things that aren't columns (node, pod IP, image) are not searched.

label:key=value keeps the objects with that label value, label:key the ones that have the label at all. Combine them with words: label:tier=web storefront.

The search box with label:tier=web, leaving the two storefront podsThe search box with label:tier=web, leaving the two storefront pods

What each pod row offers​

  • Logs, Events and, if you may, Shell (Logs, Pod shell).
  • Click the row, or Tab to the pod's name and press Enter, for its details (Detail drawers). The same works for Deployments, StatefulSets and DaemonSets.
  • The status, restarts and containers columns are explained in Pod health.
Can't see it or can't do it?
  • A tab is missing: you don't hold <kind>:list in this namespace (e.g. deployments:list).
  • No Shell button: you don't hold pods:exec, or the admin switched shells off for everyone.
  • No CPU / memory: the cluster has no metrics-server; ask your admin.
  • An empty list that should not be empty: check the search box and the namespace.