Pool Metrics

Pool metrics provide data to assess and measure performance of back end pool servers.

The Analytics information available for a pool includes Metrics related to the Pool. To access the Pool Metrics page:

1. Navigate to Applications > Pools.

2. Click a pool to open the Analytics tab.

3. Click the < icon to open the Pool Metrics tile.

The Analytics tab under a pool instance presents information about various pool performance metrics. The Data shown is filtered by the period selected. Pool Metrics Tiles continuously display time-averaged data. They govern what appears in the chart display. The following metrics are available:

Metric	Description
End to End Timing	The total time from the pool’s End to End Timing graph.
Throughput	The total bandwidth passing through the virtual service in Mbps. Pointing over this graph displays the throughput in Mbps for the selected time. Throughput is measured as bytes transferred between the client and SE. It does not include data transferred between SE and the servers.
Open Conns	The number of TCP client connections or UDP transactions currently in an open state. UDP transactions are counted even though they are technically connection-less. The number of open connections does not necessarily translate into the number of unique clients, as protocols, such as HTTP 1.1 which typically open six connections per client browser. If the Maximum Concurrent Connections setting is set for a pool server, a horizontal red bar superimposed over the chart pane will show the maximum number of connections allowed.
Estimated Capacity	Estimated capacity of the pool in terms of maximum possible connections.
Available Capacity	Available capacity of the pool in terms of currently open connections.
New Connections	The number of client connections that were completed or closed over the selected period. The metric records the number of closed connections per second.
Requests	The number of HTTP requests sent to the servers assigned to the pool. This metric also shows errors sent to servers or returned by servers. Client requests that received an error response generated by the NSX Advanced Load Balancer (such as a 500 when no servers are available) are not forwarded to the pool or tracked in this view.
Servers	Displays the number of servers in the pool and their health. The x-axis represents the number of HTTP requests or connections to the server. The y-axis represents the health score of the server. The chart enables viewing of servers in relation to their peers within the pool, thus helping to spot outliers. Within the chart pane, click and drag the mouse over server dots to select and display a table of the highlighted servers below the chart pane. The table provides more details about these servers, such as Server Name, IP Address, Port, Health, Ratio (the server’s static load-balanced ratio), and Throughput. Clicking the name of a server shows the pool’s Server Insight page, which displays additional health and resource status.
CPU	The average CPU usage during the time across all servers in the pool.
Memory	The average memory usage during the period across all servers in the pool.

Pool Metrics are classified into Layer 4 and Layer 7 metrics.

Table 1. Layer 4 Metrics

Metric	Description
l4_server.apdexc	Measures the network connection quality (errors and lossy connections) between Service Engines and servers.
l4_server.avg_available_capacity	Estimated connections per second capacity available for a server. This metric is the difference between the max capacity and the average load on a server.
l4_server.avg_bandwidth	Average transmit and receive network bandwidth between client and virtual service.
l4_server.avg_complete_conns	Rate of new connections per second.
l4_server.avg_connections_dropped	Rate of dropped connections per second.
l4_server.avg_errored_connections	Rate of total error connections per second.
l4_server.avg_est_capacity	Estimated averaged capacity of a server's connections per second summed across all SEs. Pool level metric reflects summed estimated capacity across all the servers in the pool.
l4_server.avg_goodput	Application data goodput (data excluding network headers) as bytes per second between the SE and server.
l4_server.avg_health_status	Health score status of the server. 0 is down.
l4_server.avg_lossy_connections	Rate of lossy connections per second between the SE and server.
l4_server.avg_new_established_conns	Rate of new established connections per second between SE and server.
l4_server.avg_open_conns	Number of concurrently open connections between SEs and servers.
l4_server.avg_pool_bandwidth	Transmit and receive network bandwidth between SEs and all servers in a pool.
l4_server.avg_pool_complete_conns	New connections per second across the virtual service or pool.
l4_server.avg_pool_errored_connections	Total connections classified as errored between Service Engines and all servers in a pool.
l4_server.avg_pool_new_established_conns	Total new connections per second established between Service Engines and all servers in a pool.
l4_server.avg_total_rtt	Average Round Trip Time across all completed (closed) connections.
l4_server.avg_uptime	Percent of time a server was marked as up.
l4_server.max_open_conns	Maximum number of concurrently open connections to a server.
l4_server.pct_connection_errors	Percent of network connections between Service Engines and a server that were dropped or lossy.
l4_server.pct_connection_saturation	Percent of a server's connection per second capacity that is estimated to be utilized.
l4_server.sum_connection_errors	Total number of network connections to a server that were dropped or were classified as lossy.
l4_server.sum_connections_dropped	Total number of network connections to a server that were dropped.
l4_server.sum_finished_conns	Total number of completed connections to a server.
l4_server.sum_health_check_failures	Total number of times a server was marked down by health monitors.
l4_server.sum_lossy_connections	Total number of network connections to a server that were classified as lossy.
l4_server.sum_lossy_req	Total number of HTTP requests that were classified as lossy due to high packet retransmissions.

Table 2. Layer 7 Metrics

Metric	Description
l7_server.apdexr	Quality (combination of performance and errors) of HTTP responses from servers to the virtual service.
l7_server.avg_application_response_time	Average response latency measured of pool servers.
l7_server.avg_complete_responses	Rate of server HTTP responses per second.
l7_server.avg_error_responses	Rate of HTTP error responses sent per second. Does not include errors excluded in analytics profile.
l7_server.avg_frustrated_responses	Number of HTTP requests completed which had server response latency classified as Frustrated per the virtual service analytics profile.
l7_server.avg_resp_1xx	Rate of 1xx HTTP responses sent per second.
l7_server.avg_resp_2xx	Rate of 2xx HTTP responses sent per second.
l7_server.avg_resp_3xx	Rate of 3xx HTTP responses sent per second.
l7_server.avg_resp_4xx	Rate of 4xx HTTP responses sent per second.
l7_server.avg_resp_4xx_errors	Rate of 4xx HTTP responses per second minus error codes excluded by the analytics profile.
l7_server.avg_resp_5xx	Rate of 5xx HTTP responses sent per second.
l7_server.avg_resp_5xx_errors	Rate of 5xx HTTP responses per second minus error codes excluded by the analytics profile.
l7_server.avg_resp_latency	Latency measured for pool servers.
l7_server.avg_satisfactory_responses	Number of HTTP requests completed which had server response latency classified as Satisfied per the virtual service analytics profile.
l7_server.avg_tolerated_responses	Number of HTTP requests completed which had server response latency classified as Tolerated per the virtual service analytics profile.
l7_server.avg_total_requests	Rate of HTTP requests per second received by pool servers.
l7_server.pct_response_errors	Percent of HTTP 4xx and 5xx server responses.
l7_server.sum_get_reqs	Total number of HTTP GET requests received by servers.
l7_server.sum_other_reqs	Total number of HTTP requests that are not GET or POST request received by servers.
l7_server.sum_post_reqs	Total number of HTTP POST requests received by servers.
l7_server.sum_total_responses	Total number of HTTP responses sent from servers.

Table 3. VMware Metrics

Metric	Description
vm_stats.avg_cpu_usage	Percent of server CPU used.
vm_stats.avg_cpu_wait	Percent of time virtual machine was ready to run but could not due to CPU unavailable. This could be CPU limits configured in vCenter or other virtual machines stealing CPU time.
vm_stats.avg_disk1_usage	Virtual disk1 capacity usage.
vm_stats.avg_disk2_usage	Virtual disk2 capacity usage.
vm_stats.avg_disk3_usage	Virtual disk3 capacity usage.
vm_stats.avg_disk4_usage	Virtual disk4 capacity usage.
vm_stats.avg_disk_commands_aborted	Rate of disk I/O commands that were ended prematurely in a virtual machine.
vm_stats.avg_disk_io	Rate of server disk reads plus writes per second.
vm_stats.avg_disk_read	Rate of data read from disk in kilobytes per second.
vm_stats.avg_disk_write	Rate of data written to disk in kilobytes per second.
vm_stats.avg_mem_swap_in	Total amount of data that has been read into machine memory from the swap file since the virtual machine was powered on.
vm_stats.avg_mem_swap_out	Total amount of data the VMkernel has written to the virtual machine swap file from machine memory. This statistic refers to VMkernel swapping and not to guest OS swapping.
vm_stats.avg_mem_usage	Percent of available server memory used.
vm_stats.avg_mem_vmmemctl	Virtual machine physical memory currently reclaimed from the virtual machine through ballooning. This is the amount of guest physical memory that has been allocated and pinned by the balloon driver.
vm_stats.avg_net_dropped	Rate of dropped received and transmit packets. It be an indication of network congestion.
vm_stats.avg_net_usage	Transmit plus receive network bandwidth for the virtual machine.
vm_stats.avg_port_usage	Percent of high ports used.
vm_stats.avg_uptime	Percent of time the virtual machine was up during a time interval.
vm_stats.avg_virtual_disk_commands_aborted	Number of SCSI commands that were ended prematurely.

Note:

The VMware metrics listed for Pool Metrics are also part of SE Metrics.