Edge Datapath NIC throughput high alarm on NSX Edge VM and Bare Metal Edge

Products

• VMware NSX

Issue/Introduction

Event ID: edge_health.edge_datapath_nic_throughput_high
Added in release: 3.2

Alarm Description

• Purpose: Indicates Edge Datapath NIC throughput is high
• Impact: Tx drop might be observed when throughput reaches 100%

Environment

VMware NSX-T Data Center, VMware NSX

Cause

High throughput is caused by high traffic volume. An NSX alarm is raised as a warning when the throughput is exceeding the configured limit.

Resolution

Steps to Resolve:
For 3.2.x and higher

Examine the traffic throughput levels on the Edge and determine whether configuration changes are needed. 

Edge Node Statistics

Get the CPU stats detailing the CPUs dedicated to the datapath

# get dataplane cpu stats

 

Get the memory statistics dedicated to datapath performance.

# get dataplane memory stats

 

Get the dataplane flow-cache statistics.  Flow-cache

# get dataplane flow-cache stats

 

Get a complete list of performance statistics including the CPU and Memory stats, as well as counters such as packets per second transferred, received, etc. over the time period set in <interval>.

# get dataplane perfstats <interval>

# get dataplane perfstats 10

 

To determine the throughput of an Edge for a given short period of time, run this command.

# get dataplane throughput <interval in seconds>

# get dataplane throughput 10 | json <--prints the output of this command in an easy to read json stack

 

Recommended Action:

Increased throughput support can be provided by

• Increasing Edge node form form factor size
• Add additional Edge Nodes to the cluster
• Migrate Edge Nodes to hosts that have NICs with higher bandwidth/speed. 

Maintenance window requirement for remediation? Yes

Additional Information

Starting in NSX version 3.2.4, this alarm will only trigger on bare-metal Edge's and will no longer trigger on VM Edge's.