Node-Level Monitoring

We report live per-node utilisation metrics, CPU, memory, scratch disk, IO wait, network and uptime, for every node in a cluster.

We keep the lifecycle history and the logs of each individual node alongside them.

Where node monitoring lives

1. Open Clusters from the tenant navigation panel.
2. Click a cluster name to open its detail page.
3. Open the Events tab.
4. Select the Nodes sub-tab.

note

There's no top-level Nodes tab on the cluster detail page. The cluster detail page has five tabs, Events, Configuration, Dependencies, JDBC and Access Management, and node monitoring is the Nodes sub-tab inside Events, alongside the Cluster sub-tab.

The sub-tab label carries the provisioned node count, for example Nodes 1.

Nodes sub-tab

The table has ten columns:

Column	What it shows
Node No.	The node's number. It's a link. Clicking it opens the Details for Node N dialog.
Type	A chip identifying the node's role: WORKER, or MASTER on multi-node clusters.
Status	The node's current lifecycle state, with an info tooltip.
CPU	CPU utilisation, as a mini progress bar plus a percentage.
Memory	Memory utilisation, as a mini progress bar plus a percentage.
Scratch Disk	Scratch disk utilisation, as a mini progress bar plus a percentage.
IO Wait	IO wait, as a mini progress bar plus a percentage.
Network	Two live rates — a download rate and an upload rate (for example 59.1 KB/s down and 2.2 MB/s up).
Uptime(hrs : mins)	How long the node has been up, as hrs:mins (for example 16:37).
Logs	A View logs link that opens the Node Log : N dialog.

A footer provides first/previous/next/last pagination controls and a configurable page size, defaulting to 20.

Refresh controls

The Nodes sub-tab shares the Events tab's right-aligned control row:

• Refresh Interval: a segmented control with 5s, 10s, 15s, 20s and 30s. 10s is the default.
• A Refresh button for an immediate refresh.

Node status values

Status	Meaning
STARTING	The node's instance is being launched
BOOTSTRAPPING	The node is running its bootstrap steps
RUNNING	The node is up and available to the cluster
STOPPING	The node is shutting down
STOPPED	The node is stopped
DESTROYING	The node's resources are being torn down
DESTROYED	The node has been removed
ERROR	The node failed

Details for Node N dialog

Clicking a value in the Node No. column opens the Details for Node N dialog.

The title reads Details for Node 0, followed by the node's type chip (WORKER), and the dialog header carries a refresh icon and a close icon. The body is a horizontally scrollable table with one row per node state transition:

Column	Description
Status	The state this transition entered
Start Time	When the state was entered
End Time	When the state ended
Duration	How long the state lasted
Logs	A View logs link, per transition

A row whose status is ERROR renders two extra icon buttons beside the status: a copy icon and the AssistantX icon, which runs a one-click diagnosis of the underlying error.

Node Log : N dialog

You reach this one through View logs, either from the Logs column on the Nodes table or from a row of the Details for Node N dialog, and we title it after the node whose output stream you are inspecting.

Control	Description
Title	Node Log : 0, followed by the status chip for the transition being viewed (RUNNING, ERROR, and so on). On an ERROR chip the copy and AssistantX icons appear here too.
Stdout / Stderr	A segmented toggle that switches the stream being displayed.
Sort order	An icon that flips the log between oldest-first and newest-first.
Refresh	Re-fetches the log.
Copy	Copies the displayed log.
Download	Downloads the log. Present for terminal states; it isn't shown while a log is still streaming.
Log body	The log lines, with line numbers and a floating scroll-to-bottom button.
System Error Messages	A collapsible panel at the bottom, with its own copy icon, holding the system-level error messages for the node.
Close	Dismisses the dialog.

Cluster-level events

The Cluster sub-tab, next to Nodes, carries the cluster's own lifecycle events, newest first: the status, who triggered the transition, the start and end times, and the duration humanised as 16h 36m or 34s. Its Triggered By column shows the email address of whoever initiated the transition, or -- when the platform initiated it, and the statuses it walks through are INITIATING, RUNNING, STOPPING, STOPPED, DESTROYING, DESTROYED and ERROR. Both sub-tabs share the same right-aligned control row, so the 10s refresh interval selected on Nodes also determines how frequently Cluster re-polls. See Monitoring.