Some network metrics are considered Key Performance Indicators (KPIs) because they are essential measures of whether the network or a specific part of the network, is performing successfully. Network Agents collect network KPIs for all monitored tiers, nodes, network links, and application flows. You can view KPI metrics for these elements:

View Network KPIs

For the Entire Application

The Network Dashboard includes Throughput, TCP Loss, and PIE charts that measure network health for the entire application.

  • Network DashboardThe flow map shows KPIs for each link. Click a link to view KPI time charts and KPIs per Connection.
  • Right-click dashboardRight-click a link in the Network Dashboard and select View Metrics. See KPI Metrics in Right-Click Dashboards

For a Tier

  • Network Flow MapClick a tier to view KPI time charts.
  • Right-click dashboard – Right-click a tier in the Network Dashboard and select View Metrics. See KPI Metrics in Right-Click Dashboards

For a Node

  • Tiers & Nodes page – Set the Show Data menu to Network. The table shows network KPIs and the Network Agent status for each node.
  • Node Dashboard – Right-click the node and select View Metrics. See KPI Metrics in Right-Click Dashboards

For a Connection

Network Agents do not collect metrics for Connections by default. To collect metrics for the Connections used by a specific node, change the Monitoring Mode on the Agent. See Dynamic Monitoring Mode and Network Visibility

Connections Explorer – Access the Network Dashboard and select Connections. The Explorer shows network KPIs for all network links. Drill down into a network link to view the Connections used on that link. If a Network Agent has collected metrics for a Connection, you can:


KPI Metric Descriptions

MetricDescriptionDefault Monitoring Mode
Throughput

The average rate of Application X traffic sent and received over one Connection, or a set of Connections.

Correlated spikes along with Response Time indicate bottlenecks.

KPI
Latency

The average TCP Latency (round-trip time) for a packet to go from one application tier the other, and back again. When TCP effects are not significant, Latency primarily measures network round trip times. Network latency is primarily a factor of the physical distance between two nodes and remains consistent over time. 

Correlated spikes along with Response Time indicate that TCP effects between two tiers (such as packet queueing and delayed acknowledgments) are affecting application performance. 

KPI
TCP Loss (mille)

The average number of packets (per 1000) that were sent but never received. Packet loss can degrade application performance significantly. TCP detects and re-transmits lost packets; this ensures reliable transmission but introduces delays.

  • Correlated spikes with Response Time indicates that network congestion between the sender and receiver is affecting application performance. 
  • Correlated spikes with Throughput indicates that a tier is sending more data than the receiving tier or the network can handle.
KPI
Errors
The number of TCP messages sent indicating an error in setting up the connection (SYN errors) or tearing down the connection (FIN errors) for a specific TCP session.  KPI
PIE (Performance Impacting Events)

The rate of PIE (Performance Impacting Events), which indicate performance issues on one node, both nodes, or the network between two nodes:

KPI