Terminal UI

Cluster Overview

Overview shows cluster health, resource usage, warnings, and inventory. Select a cluster connection and press   11   or   :1   to open it.

It covers the selected cluster across all namespaces. Selecting a workspace instead shows its saved connections.

Take a first look

Section

Look for

Open with Enter

Cluster Pulse

CPU, memory, Node health, Pods, restarts, autoscaling

A chart or resource list

Recent Deployments

Reported rollout and release changes

The change record

Needs Attention

Failures and warnings

The resource or issue group

Infrastructure Snapshot

Kubernetes and detected integration inventory

Matching objects

For an investigation, start with Needs Attention. Open a resource, then its Events or logs.

Cluster Pulse

CPU and memory compare measured usage with Node allocatable capacity. Requests and limits are shown separately because they describe resource budgets, not consumption.

Press Enter on CPU or Memory for a larger chart. Enter again opens Source details with API sources, timestamps, and calculation notes. Esc returns through the chart to Overview.

Charts keep up to 15 minutes of samples. Gaps longer than 30 seconds remain visible. Press   e   to expand Pulse,   /   to find text, or   ff   for more room.

Label

Meaning

Unavailable

Metrics are missing, incomplete, or do not match the observed Nodes

STALE

The source measurements have expired

Requests / limits

Declared budgets, including active unscheduled Pods

Node pressure

Nodes reporting pressure conditions

Pod phases

Reported phases, including completed Pods

Restarts

Cumulative restart counts

HPA direction

Desired replicas compared with current replicas

Metrics require access to   metrics.k8s.io  . Missing measurements are not zero. Plous does not provide network throughput without a telemetry source.

Understand change indicators

Restart and replica-change indicators compare reads of the same resource identities. Missing data, replacement resources, counter resets, or gaps over 45 seconds prevent a comparison.

An absent indicator does not confirm that nothing changed.

Recent Deployments

This feed shows reported Deployment rollouts, Helm revisions, Argo CD sync operations, and Flux condition changes. It uses their reported timestamps.

The current cluster session keeps up to 200 distinct records. History marks older observations. Enter opens the saved observation rather than claiming it is current.

For saved workspace changes and recorded Helm operations, use Activity. Neither view is a complete cluster audit log.

Needs Attention

Errors use × and warnings use !. Rows are ordered by severity and reported change time. Pods with the same controller and cause can form a group. Enter opens its members.

NEW means an error appeared between comparable complete reads within 45 seconds. It does not identify when the incident began.

An empty list means no warnings were found in the data Plous could read. Check availability before treating that as a healthy cluster.

Infrastructure Snapshot

Select a tool card, choose a counter with Up / Down, and press Enter for matching objects. Esc restores the card.

The Availability summary explains missing APIs, denied access, unavailable metrics, and incomplete results.   ?   means unknown.   +   and Partial mark incomplete counts.

Argo CD Applications stored on another management cluster are outside this cluster view.

Arrows follow the displayed layout. Use   /   to filter lower lists,   e   to expand a section, and   ff   to expand main. Alt+Left / Alt+Right changes panels.

Selecting an Overview resource preserves the filters and namespace saved in Explore. Esc returns to the selected Overview item.

Refresh, freshness, and retention

Data

Automatic refresh

Retention

CPU and memory

Every 5 seconds

Metrics expire after 90 seconds. Up to 15 minutes of samples in memory.

Resource and integration inventory

Every 15 seconds

Marked STALE after 45 seconds

Recent changes

With inventory reads

Up to 200 records in this cluster session

Press   g   to refresh explicitly. Reads have a 10-second deadline. A repeated fetch may contain the same measurement.

Overview does not collect metrics after you leave or save historical charts. An assistant can save a separate read as Sources.

Next steps

Try Investigate a restarting Pod or read the Kubernetes guide.

View source

Plous · Open source · Early development

Search Plous docs

Search the documentation.

Documentation versions