SE Metrics

SE metrics includes information on connection, bandwidth, transaction, and memory usage for an SE.

The Analytics information available for an SE includes performance metrics for the SE over a defined time.

To access the SE Metrics page:

Navigate to Infrastructure > Cloud Resources > Service Engine.
Select a cloud from the Select Cloud drop-down menu.
Click the name of an SE to open the Analytics tab for the SE.
Click the < icon to open the Service Engine Metrics tile.

The Analytics tab of the SE presents performance metrics information for the SE. Data shown is filtered by the time selected. Service Engine Metrics Tiles continuously display time-averaged data. They govern what appears in the chart display. The following metrics are available:


Metric	Description
Throughput	Total bandwidth flowing through the SE for all virtual services being hosted by that SE. It includes the bandwidth flowing in and out of the SE between the client and the virtual service, and the traffic between the SE and the servers. An SE can report approximately double the throughput of its virtual services.
CPU Usage	Displays the utilization of the CPUs allocated to the SE. The total number of CPUs appears in the SE Quick Info pop-up window. Under normal conditions, CPU usage must not regularly exceed 90%, as this can cause latency in the virtual services and disrupt the client experience. The CPU Usage metric tile shows a horizontal bar indicating current usage with a red line at the right to indicate how close the SE is pushing the limits of its available CPU capacity. The CPU usage can be indirectly controlled or improved by taking actions, such as: Configuration: Changing the configuration of virtual services, such as changing SSL or compression settings, impacts the CPU usage. CPU Allocation: Allocating more vCPUs per SE. The default setting is two vCPUs per SE. Increasing this number is particularly useful for tasks such as SSL termination or compression, which heavily consume CPU resources. The setting for the number of vCPUs assigned to an SE is in the SE group. Scale Out: Reduce the CPU load by scaling the SE’s virtual services across additional SEs. This increases the total capacity and reduces the load on this SE. The high availability setting of the SE group dictates when a virtual service should be scaled out across additional SEs or simply migrated away from a busy SE. CPU Reservation: By default, CPUs resource is not reserved in a VMware deployment. Within vCenter, you can enable reservation for the SE’s virtual machine. This guarantees that other virtual machines sharing the same physical host server are not able to borrow or compete for CPU resources. This setting can be changed in the SE group properties. Changes will apply for only new SEs. Changes for the existing SEs have to be done manually within vCenter.
Memory Usage	Amount of used versus available memory. Memory utilization should not exceed 90% for an extended period.
Rx Packets	The received network packets for the SE.
Tx Packets	The number of packets successfully transmitted for the SE.
Interface Throughput	The combined throughput for all network interfaces utilized by this SE. Throughput is measured as both client and server side of any virtual services, plus the management traffic between the SE and the Controllers.
Virtual Service Throughput	The combined throughput for all network interfaces utilized by this SE. Throughput is measured as both client and server side of any virtual services, plus the management traffic between the SE and the Controllers.
Connection Memory Usage	Percentage of connection memory used by the SE.
Dynamic Memory Usage	Percentage of dynamic memory used by the SE.
SSL Cache Usage	Percentage of SSL cache memory used by the SE.
Persistent Table Usage	Persistent Table used by the SE expressed as %. Since client IP persistence is stored locally on each SE, larger tables consume more memory.
Buffer Usage	Percentage of buffer memory used by the SE.

The list of SE metrics available in the NSX Advanced Load Balancer system are listed in the table below:


Metric	Description
`se_if.avg_bandwidth`	Transmit and receive network bandwidth across all SE interfaces.
`se_stats.avg_connection_mem_usage`	Percent of connection memory consumed.
`se_stats.avg_connections`	Rate of client network connection attempts (SYNs) per second across all virtual services for an SE.
`se_stats.avg_connections_dropped`	Number of connections dropped or failed to establish across all virtual services for an SE. Excludes drops due to a policy action.
`se_stats.avg_cpu_usage`	Hosts view of the Service Engines actively utilized CPU usage as a percent of total available CPU.
`se_stats.avg_disk1_usage`	Percent of SE disk 1 capacity used.
`se_if.avg_eth0_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth10_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth11_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth12_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth13_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth14_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth15_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth16_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth17_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth18_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth19_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth1_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth20_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth21_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth22_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth23_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth2_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth3_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth4_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth5_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth6_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth7_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth8_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_if.avg_eth9_bandwidth`	Transmit and receive network bandwidth for an SE interface.
`se_stats.avg_mem_usage`	Percent of allocated memory used.
`se_stats.avg_packet_buffer_header_usage`	Percent of all network packet buffers used.
`se_stats.avg_packet_buffer_large_usage`	Percent of large network packet buffers used.
`se_stats.avg_packet_buffer_small_usage`	Percent of small network packet buffers used.
`se_stats.avg_packet_buffer_usage`	Percent of total configured network packet buffers used.
`se_stats.avg_persistent_table_usage`	Percent of session persistence table used.
`se_if.avg_rx_bandwidth`	Received network bandwidth across all interfaces for an SE.
`se_if.avg_rx_bytes_dropped`	Bytes per second of received packets that were dropped. Includes packets across all SE interfaces.
`se_if.avg_rx_pkts`	Received network packets across all SE interfaces.
`se_if.avg_rx_pkts_dropped`	Received packets dropped per second. Includes packet drops across all interfaces.
`se_stats.avg_ssl_session_cache_usage`	Percent of SSL session cache used.
`se_if.avg_tx_pkts`	Network packets transmitted across all SE interfaces.
`se_if.max_se_bandwidth`	Maximum bandwidth through an SE within the sample period.
`se_stats.pct_connections_dropped`	Percent of SE connections dropped.
`se_if.pct_rx_bytes_dropped`	Percent of SE bytes dropped.
`se_if.pct_rx_pkts_dropped`	Percent of SE packets dropped.
`se_stats.pct_syn_cache_usage`	Percent of SYN cache used. Higher usage indicates too many incomplete open connection attempts.
`se_stats.avg_dynamic_mem_usage`	Percent of average dynamic memory used.

For more information on VMware metrics related to SE, see VMware Metrics.