Edge memory usage very high alarm on NSX-T Edge Node due to RCPM memory exhaustion

Products

• VMware NSX

Issue/Introduction

This article describes the causes and remediation steps for the "Edge memory usage very high" fatal alarm in VMware NSX-T environments, specifically when triggered by routing control plane service exhaustion.

Symptoms:

• The NSX Manager UI reports a FATAL alarms:  edge_memory_usage_high and edge_memory_usage_very_high.
• A bgp_down event or BGP flap is observed in syslogs prior to the alarm.
• Edge logs display error codes: MP701099 correlates the bgp_down event with the subsequent memory exhaustion.

2026-07-30T09:05:19.760Z v##-###n-0#.####z.###c.##e NSX 76331 MONITORING [nsx@6876 alarmId="2#######-4###-###a-9###-###########0" alarmState="RESOLVED" comp="nsx-manager" entId="5#######-###3-###a-b###-###########d" errorCode="MP701099" eventFeatureName="routing" eventSev="HIGH" eventState="Off" eventType="bgp_down" level="ERROR" nodeId="4#######-###8-###d-9###-###########3" subcomp="monitoring"] User resolved.

2026-07-30T09:05:19.798Z v##-###n-0#.####z.###c.##e NSX 76331 MONITORING [nsx@6876 alarmId="8#######-c###-###e-a###-2###########" alarmState="RESOLVED" comp="nsx-manager" entId="2#######-###b-4###-###9-###########7" errorCode="MP701099" eventFeatureName="edge_health" eventSev="CRITICAL" eventState="Off" eventType="edge_memory_usage_very_high" level="FATAL" nodeId="2#######-b###-###2-a###-3###########" subcomp="monitoring"]  message repeated 2 times: [User resolved.]

• CLI output from top shows the rcpm service consuming a disproportionate amount of resident memory (e.g., >35% of the total Edge node memory).

Environment

VMware NSX

Cause

• RCPM high being noticed right after a BGP flap or down event, which is responsible for handling routing and control plane protocols.
• A BGP flap or connectivity event triggers a memory exhaustion state within the rcpm (Routing Control Plane Manager) daemon.
• The daemon fails to reclaim memory during message cloning or mempool creation, leading to a fatal memory utilization threshold.

Resolution

1. Enter Maintenance Mode: Immediately place the impacted NSX-T Edge Node into maintenance mode to gracefully shift active workloads and prevent traffic disruption.
2. Restart Edge Node: Once maintenance mode is confirmed, restart the Edge Node to terminate the rcpm process and reclaim memory.
   • WARNING: Restarting without maintenance mode may cause active traffic loss.
3. Verify Stability: Post-reboot, verify memory levels using the 'top' command from root mode.

Additional Information

High memory utilization in rcpm maps to documented NSX error states EDG0200010 (Mempool creation failed in RCPM) and EDG0200017 (Memory issue cloning messages sent by NestDB to RCPM).

KB: NSX Error Codes.