NodePools and Cluster Autoscaler

To view the dashboard, go to the NodePools and Cluster Autoscaler dashboard.

For instructions about accessing CoreWeave Grafana dashboards, see Access and use CoreWeave Grafana dashboards.

The NodePools and Cluster Autoscaler dashboard gives you a single place to inspect the state of your CKS Node Pools and the Kubernetes Cluster Autoscaler that scales them. Use it to review Node Pool capacity and conditions, read Node Pool events, and track autoscaler scaling activity, errors, and logs without querying metrics or logs directly. For background on how autoscaling works and how to configure it, see Autoscale Node Pools.

Node Pools overview

The NodePools Overview table provides a global view of all your Node Pools. It renders the spec fields, status fields, and conditions for each Node Pool in one place. For details about what each column represents, see the Node Pool reference documentation.

Column	Description
NodePool	The name of the Node Pool.
Org ID	The organization that owns the Node Pool.
Cluster	The cluster the Node Pool belongs to.
Zone	The availability zone where the Node Pool runs.
Instance Type	The instance type provisioned by the Node Pool.
Compute Class	The compute class assigned to the Node Pool.
Target Nodes	The number of Nodes the Node Pool is requesting.
Current Nodes	The number of Nodes present in the cluster for the Node Pool.
Progressing Nodes	The number of Nodes that have been assigned but have not yet fully booted into the cluster.
Queued Nodes	The number of queued Nodes waiting to be assigned.
Prefill Nodes	The number of Nodes in the Node Pool prefill flow.
Autoscaling	Whether autoscaling is enabled on the Node Pool.
Validated	The `Validated` condition, which reports whether the Node Pool spec passed validation.
AtTarget	The `AtTarget` condition, which reports whether the Node Pool has reached its target Node count.
Capacity	The `Capacity` condition, which reports whether the zone has capacity to provision the requested Nodes.
Quota	The `Quota` condition, which reports whether the request is within the organization’s quota.
ReconfigRequired	The `NodeReconfigurationRequired` condition, which reports whether Nodes need to be reconfigured to match the spec.

Node Pool events

The NodePool Events table renders the events recorded for your Node Pools, such as scaling decisions and quota warnings. Filter the table by event type (Normal or Warning) and event reason to focus on specific activity. Because events come from exported logs, set the Log Source variable to a source that matches the Region and Zone you have selected. Otherwise, the table cannot render events.

Cluster autoscaler

The Cluster Autoscaler section reports the activity and health of the Cluster Autoscaler. To drill into a single cluster, set the Org ID and Cluster variables at the top of the dashboard.

Cluster autoscaler metrics

These panels summarize how the autoscaler is scaling your cluster and how its main loop is performing. The metrics panels use the Data Source, Region, Zone, Org ID, and Cluster variables.

Panel	Description
Safe to Autoscale	Whether the cluster is healthy enough for autoscaling.
Autoscaled NodePools Count	The number of Node Pools managed by the Cluster Autoscaler.
Autoscale Max Nodes Count	The maximum number of Nodes aggregated across all autoscaled Node Pools.
Autoscaled Nodes Count	The number of Nodes in the cluster managed by the autoscaler.
Cluster Autoscaler Duration p95 By Function	The time taken by various parts of the autoscaler main loop.
Errors	The number of autoscaler loops that failed due to an error.
Failed Scale Ups	The number of times a scale-up operation has failed.
Unneeded Nodes	The number of Nodes currently considered unneeded by the autoscaler.
Scaled Down Nodes	The number of Nodes removed by the autoscaler.
Unschedulable Pods Count	The number of unschedulable (`Pending`) Pods in the cluster.
Scaled Up Nodes	The number of Nodes added by the autoscaler.

Cluster autoscaler logs

The Logs panel shows exported logs from the Cluster Autoscaler. The logs panel uses the Log Source, Region, Zone, Org ID, and Cluster variables. As with Node Pool events, set the Log Source variable to a source that matches the Region and Zone you have selected so the logs render correctly.

Dashboard variables

The variables at the top of the dashboard filter the panels below them. Not every panel uses every variable.

Variable	Description
Data Source	The metrics data source for the metric panels.
Log Source	The log data source for the events and logs panels. Match this to your selected region and zone.
Region	Filter to a specific region.
Zone	Filter to a specific availability zone.
Org ID	Filter to a specific organization.
Cluster	Filter to a specific cluster.
Instance Type	Filter to Node Pools of a specific instance type.
Compute Class	Filter to Node Pools of a specific compute class.
NodePool	Filter to a specific Node Pool.
Event Type	Filter Node Pool events by type (`Normal` or `Warning`).
Event Reason	Filter Node Pool events by reason.

​Node Pools overview

​Node Pool events

​Cluster autoscaler

​Cluster autoscaler metrics

​Cluster autoscaler logs

​Dashboard variables

Node Pools overview

Node Pool events

Cluster autoscaler

Cluster autoscaler metrics

Cluster autoscaler logs

Dashboard variables