Unmanaged or Down Node Displays Stale CPU Data in "Nodes with High CPU Load" or Top XX Widgets

8/3/2026 10:47 PM

8/3/2026 10:47 PM

Environment

• SolarWinds Platform (all versions)

• Network Performance Monitor (NPM)

• Any built-in Top XX or "Nodes with High Average CPU Load" widget (ResourceID 780)

Symptoms

• A node that is powered off, Down, or unmanaged continues to appear in the "Nodes with High Average CPU Load" widget with its last polled CPU percentage (e.g., 99%)

• The node displays the Unmanaged status icon (blue circle) or Down status icon (red) in the widget

• The stale value persists indefinitely until the node is either re-managed and re-polled, or deleted from SolarWinds

Network Performance Monitor

Two platform behaviors combine to produce this issue:

1. The CPULoad column retains the last polled value indefinitely.

When a node is unmanaged, polling stops but the CPULoad value in the Nodes/NodesData table is NOT reset to 0 or NULL. The SolarWinds Platform preserves all configuration, entity relationships, pollers, custom properties, and historical statistics when a node is unmanaged — nothing is deleted or modified beyond the unmanage flag itself. The official KB article confirms: "Polling jobs are suspended, so no further data is collected."

Reference: What happens when I unmanage a node in the SolarWinds Platform

2. The built-in "Nodes with High Average CPU Load" widget does not exclude unmanaged or Down nodes.

The SWQL query used by the OOTB widget (ResourceID 780, source method TopXXSWQLDAL.GetCPULoad) contains only:

WHERE CPULoad >= 0
ORDER BY CPULoad DESC

There is no filter for UnManaged = FALSE or Status != Unmanaged/Down. Any node with a CPULoad value >= 0 will appear in the widget regardless of its managed state or operational status.

Option A — Edit the widget to exclude unmanaged and Down nodes (recommended, no downtime required)

1. Navigate to the dashboard view containing the "Nodes with High Average CPU Load" widget

2. Click the Edit button on the widget

3. Add a filter condition to exclude unmanaged nodes:

   • Filter: "Unmanaged" → "is equal to" → "No"

4. Optionally add a second filter to exclude Down nodes:

   • Filter: "Status" → "is not equal to" → "Down"

5. Save the widget

After applying the filter, unmanaged and Down nodes will no longer appear in the widget even if they retain stale CPU values in the database.

Option B — Delete the node from SolarWinds (no downtime required)

If the server is permanently decommissioned:

1. Navigate to the node's details page or Manage Nodes

2. Delete the node from SolarWinds

3. Wait for DB Maintenance to run (default schedule: daily) — the orphaned data will be automatically cleaned up

After deletion and DB Maintenance, the node will no longer appear in any widget.

Option C — Re-manage and re-poll the node (no downtime required)

If the server is back online and should be actively monitored:

1. Navigate to Manage Nodes, select the node, and click Manage Again from the More Actions menu

2. Wait for the next polling cycle (default: 10 minutes for Statistics Poll Interval)

3. The CPULoad value will update to the current actual CPU utilization

Notes

• This behavior applies to all Top XX widgets that read current CPULoad, PercentMemoryUsed, ResponseTime, or PercentLoss from the Nodes entity — not just the CPU widget

• DB Maintenance does NOT clear stale CPULoad values on unmanaged nodes — it only removes orphaned data from deleted nodes

• If using custom SWQL-based resources or reports, add WHERE UnManaged = FALSE to exclude unmanaged nodes from results