Overloaded storage alarm in DynaTrace tool.

Products

• VMware vSphere ESXi

Issue/Introduction

•  DynaTrace monitoring tool triggers an 'Overloaded storage' alarm on an ESXi host.

• ESXi performance metrics appear normal in vCenter; no slowness or latency is observed on running VMs.

 

 

Environment

VMware vSphere ESXi 8.x
VMware vSphere ESXi 9.x

DynaTrace Monitoring Tool

Cause

DynaTrace triggers this alarm when storage metrics exceed defined anomaly detection thresholds i.e when command aborts are higher than X in 3 out of 5 samples In many cases, these metrics are inflated because the ESXi host is attempting to perform I/O on storage devices that belong to a decommissioned storage array but not cleaned up on the ESXi host.

• When a LUN is unmapped or deleted from a storage array before being administratively detached from the ESXi host, the host attempts to probe for the paths which is no longer exists . This leads to aborting IO requests with SCSI sense code H:0x5.
  
  2026-07-20T03: 56:05. 602Z Wa (180) vmkwarning:### 2097994) WARNING: NMP: nmpDeviceAttemptFailover: 718: Retry world failover device naa.########################## - failed to issue command due to Not found (APD), try again ...
2026-07-20T03:56:05.6022 Wa (180) vmkwarning:### 2097994) WARNING: NMP: nmpDeviceAttemptFailover: 770: Logical device "naa.##########################": awaiting fast path state update.
2026-07-20T03:56:14.599Z In (182) vmkernel: ###:2097718) ScsiDeviceIO: 4670: Cmd (0x45d9955e79c0) 0x12, cmdId. initiator=0x43094d426590 CmdSN 0x101158e from world 0 to dev "naa.##########################" failed H:0x5 D: 0x0 P:0x0 Cancelled from NMP layer. Cmd
2026-07-20T03:56:14.599Z(180) vmkwarning: ####:2097718) WARNING: NMP: nmp_DeviceStartLoop: 790: NMP Device "naa.########################## is blocked.Not starting I/0 from device.
2026-07-20T03:56:14.599Z In(182) vmkernel: ###:2097697) ScsiVmas: 1094: Inquiry for VPD page 00 to device naa.########################## with error Timeout

• Storage devices are reported to be having dead paths with unavailable targets.

naa.########################## : EMC Fibre Channel Disk (naa.##########################)
vmhba2 : C0 : T0 : L1 LUN:1 state:dead fc Adapter: Unavailable Target: Unavailable
vmhba2 : C0 : T1 : L1 LUN:1 state:dead fc Adapter: Unavailable Target: Unavailable
vmhba64 : C0 : T3 : L1 LUN:1 state:dead fc Adapter: Unavailable Target: Unavailable
vmhba 64 : C0 : T4 : L1 LUN:1 state:dead fc Adapter: Unavailable Target: Unavailable

 

• Operational status of the storage device shows as  "Dead or Error" on the ESXi host.

 

Resolution

Follow these steps to clean up decommissioned storage devices on the ESXi host:

1. Migrate Virtual Machines: Move all VMs to a different host within the cluster to ensure no downtime.
2. Reboot the Host: Perform a host reboot to clear stale storage device entries.
3. Rescan Storage: After the host is back online, perform a cluster-level storage rescan:
   • Right-click the Cluster.
   • Select Storage > Rescan Storage.
   • Ensure both Scan for new storage devices and Scan for new VMFS volumes are selected.

If the alarm persists, engage the DynaTrace vendor to investigate the monitoring thresholds.

Additional Information

For more details refer to the DynaTrace anomaly detection schema documentation.

Anomaly detection for VMware

For further guidance on contacting support, please see Contact Broadcom support