ON THIS PAGE
 

Traffic Analytics Report

This topic provides an overview of the traffic reports you can generate in the HPE Networking Apstra Data Center Director GUI. The traffic analytics report analyzes device traffic patterns and trends in the data center. To learn how to generate this report, see Generate Analytics Report.

Note: This topic shows examples of reports for hardware models and traffic analysis for spine switches. You can also generate similar reports for leaf switches.

Data Center Overview

This section provides reports on deployed hardware models in the data center. It also shows a chart of traffic rates for server-facing and fabric-facing interfaces on leaf switches.

Device Hardware Model and Operation System Inventory Overview

Figure 1 shows a summary of hardware models and corresponding operating systems for all devices deployed in the current data center.

Figure 1: Summary of Hardware Models Circular chart comparing QFX10002-36Q and VIRTUAL-EX9214. Both have a count of 2 with versions 23.4R2-S4.11 and 23.4R2-S4.7 respectively.

Aggregated Server-Facing vs. Fabric-Facing Traffic Rate

Interfaces on leaf switches can be either server-facing or fabric-facing.

  • Server-facing interfaces connect to anything that's not a spine switch (hardware server, generic system and so on). Server-facing traffic is the sum of both the transmit (TX) and receive (RX) traffic rates for all server-facing interfaces on all leaf switches.

  • Fabric-facing interfaces connect to spine switches. Fabric-facing interfaces connect to spine switches. Fabric-facing traffic is the sum of both TX and RX traffic rates for all spine-facing interfaces on all leaf switches.

Figure 2: Server-Facing vs. Fabric-Facing Traffic Bar graph showing network traffic utilization from April 21 to April 27, 2025. Green bars represent server-facing traffic and blue bars represent fabric-facing traffic. Fabric-facing traffic is consistently higher than server-facing traffic across the time period.

Traffic Analysis for Spine Switches

The following section includes detailed analysis of spine switches.

Spine Switch Inventories

The spine switch inventory lists the number of spine devices in the data center. Figure 3 provides details for a each switch, including the switch hardware model, system ID, pod, hardware model, OS type, management IP, and OS version. These details are collected dynamically from the graph and from various device information collectors.

Figure 3: Spine Switch Inventory Network device details table: label bp1-spine1, system_id 52540052DB73, pod pod1, vendor Juniper, hw_model VIRTUAL-EX9214, os_family Junos, management_ip 10.28.35.19, device_os_version 23.4R2-S4.7.

Fabric Utilization Aggregated View

This section shows the aggregated traffic rate, IQR traffic rate, and peak RX utilization for spine switches. Fabric utilization measures bandwidth usage in a switch's internal fabric. This metric indicates data flow efficiency through the switch and its components.

Spine Aggregated Traffic Rate

Figure 4 shows the formulas used to calculate the utilization (in percentages) of spine switches. Figure 5 shows a chart of all spine switch utilizations.

Figure 4: Spine Switch Calculation Formulas for network traffic and capacity: Total RX Traffic Rate is the sum of RX traffic rates of all interfaces. Total Capacity is the sum of speeds of all interfaces. Utilization is the ratio of Total RX Traffic Rate to Total Capacity.
Figure 5: Spine Switch Utilization Line graph of bp1-spine1 52540052DB73 pod1 utilization from April 21-27, 2025, showing a consistent 0 percent throughout.

Spine Aggregated Utilization IQR

Interquartile range (IQR) measures statistical dispersion (spread of data). Traffic rates vary over time. An IQR analysis helps identify a normal range of values and how often extreme values are observed.

Figure 6 shows the formulas used to calculate IQR-related thresholds. Q1, Q2 and Q3 are used to render the box in the middle of the plot. The lower and upper fence values are used to render the whiskers.

Figure 7 shows the traffic rate IQR for all spine switches during the query time window.

Figure 6: IQR Formula Definitions and formulas for statistical quartiles and fences: Q0 minimum value; Q1 lower quartile; Q2 median; Q3 upper quartile; Q4 maximum value; IQR is Q3 minus Q1; Lower Fence is max of Q0 and Q1 minus 1.5 times IQR; Higher Fence is min of Q4 and Q3 plus 1.5 times IQR. Used to identify outliers.
Figure 7: Traffic Rate IQR Chart TX utilization graph with System ID bp1-spine1 52540052DB73 on x-axis shows 0 percent utilization indicating minimal transmission activity.

Spine Peak RX/TX Utilization

Bursty traffic can cause issues even when the average traffic rate is acceptable. Peak RX/TX traffic rates are monitored on all deployed switch interfaces during each aggregation period.

Spine Peak RX Traffic Rate

Figure 8 compares peak RX traffic rates observed on all spine switches during the query time window.

Figure 8: Peak RX Traffic Rate Peak TX utilization graph for device bp1-spine1 52540052DB73 in pod1 from April 21 to 27, 2025 shows 0 percent utilization.

Spine Peak Interface RX Utilization IQR

The following chart shows the peak interface RX traffic rate IQR for all spine switches.

Figure 9: Peak Interface RX Utilization IQR RX utilization graph for system ID bp1-spine1 52540052DB73 pod1 showing 0 percent utilization indicating minimal activity.

Spine Peak TX Utilization

The following chart shows the peak interface TX traffic rate for all spine switches.

Figure 10: Peak TX Traffic Rate Peak TX utilization graph for device bp1-spine1 52540052DB73 in pod1 from April 21 to 27, 2025 shows 0 percent utilization.

Spine Peak Interface TX Utilization IQR

The following chart shows the peak interface TX traffic rate IQR for all spine switches.

Figure 11: Peak Interface TX Utilization IQR TX utilization graph with System ID bp1-spine1 52540052DB73 on x-axis shows 0 percent utilization indicating minimal transmission activity.

Fabric Utilization Individualized View

You can also view reports that show detailed traffic analysis examples for each individual spine switch. The reports include examples of fabric utilization trends, CPU and memory usage, and RX/TX utilizations as shown in the following charts.

Figure 12: Fabric Utilization and Trend Analysis Line graph showing fabric utilization at 0 percent from February 25 2025 to March 4 2025 with no change over time.
Figure 13: CPU and Memory Usage Line graph showing CPU and memory utilization from April 21 to April 27, 2025. CPU is near 0 percent, memory is constant at 45 percent.
Figure 14: RX Utilization Line graph of RX Utilization in bps from April 21 to April 27, 2025. Shows utilization for six connections: green line for ge-0/0/1->bp1-dual-leaf1 EP202, blue line for ge-0/0/3->bp1-dual-leaf2 EY304, purple line for ge-0/0/2->bp1-dual-leaf2 EY304, orange line for ge-0/0/5->bp1-single-leaf3 52540072BE8D, brown line for ge-0/0/4->bp1-single-leaf3 52540072BE8D, and red line for ge-0/0/0->bp1-dual-leaf1 EP202.
Figure 15: TX Utilization User interface of a network tool showing Fabric Settings in the Staged tab under Fabric Policy. MTU settings include Fabric MTU 9170, Default IP Links MTU 9020, and Default SVI MTU 9000. Routing Zone Footprint Optimization is enabled. Virtual MAC Address is 00:1c:73:00:00:01. Modify Settings button available. Tabs like Dashboard, Anomalies, Analytics, Staged, Uncommitted, Active, and Time Voyager are at the top.

Related Documentation