All clusters

All clusters

All clusters is an overview of several clusters on one page: a health card for each cluster you choose, tables of pods, workloads and Warning events across them, and a search by name in all of them. It is read-only, and it connects a cluster only when you click Connect.

Contents: Opening it · Choosing clusters · Connecting · Health cards · Tables across clusters · Find by name · Limits

Opening it

Click All clusters at the top of the cluster rail. The page takes the main area; the cluster you had open steps back but stays connected. Click All clusters again, or open any cluster (from the rail, the command palette, a notification or a link on the page), to leave the overview.

Choosing clusters

Under Clusters to show, tick the clusters the overview should show, up to 10. The choice is saved in the settings and kept between runs. Choosing a cluster does not connect it.

Connecting

Opening the page never connects anything, because a connection can start a sign-in through an exec plugin or a cloud CLI (aws, gcloud, az, kubelogin…). A chosen cluster that is not connected shows Not connected. and a Connect button; Connect all selected connects every chosen cluster, two at a time (the rest show Waiting to connect…). A failed connection shows its error on the card, with Reconnect. Clusters you connected elsewhere in Sextant show up connected here as well.

Health cards

For each connected cluster the card shows:

• Version: the kubelet versions of its nodes (a range when they differ);
• Nodes: how many are ready;
• CPU and memory in use against the nodes' allocatable capacity, when metrics are available;
• Problems by category, by the same rules as the Problems page (nodes, pods, workloads, jobs, volume claims, certificates), and the number of Warning events;
• Installed tools: Flux, Argo CD and Gateway API, when the cluster has them, and whether they have problems.

The card's open button opens the cluster. The opt-in Security category (Security posture) is not shown on the cards.

Tables across clusters

The tabs below the cards list objects of every connected chosen cluster, with a Cluster column:

Tab	What it lists
Pods	Namespace, name, status, ready containers, restarts, node, age
Workloads	Deployments, Stateful Sets and Daemon Sets: namespace, kind, name, ready replicas, age
Warning events	The newest Warning events (up to 200 per cluster): namespace, object, reason, message, age

Filter by Cluster and Namespace. Each table shows up to 200 rows per cluster after the filters and says how many more there are; narrow the filters to see them. Click a row to open the object in its cluster.

Find by name

Find by name searches every connected chosen cluster at once, the way the command palette does: by default in Pods, Deployments, Stateful Sets, Daemon Sets, Services and Ingresses; start with a kind to search another one (cm web, secret db). Nothing is searched before you type, and only names, namespaces and labels are read. Results are grouped by cluster and kind; click one to open it in its cluster.

Limits

• The overview reads only: there are no actions on the page. Open an object to act on it in its cluster.
• At most 10 clusters are shown, and Sextant keeps at most 5000 objects of a kind per cluster for the tables; the page says when a cluster has more.
• The overview shares its watches with the Problems page and notifications, so opening it adds little load on clusters that are already watched.
• AI agents and extensions get nothing new from it.

Next: Getting started, Problems and notifications.