CCNA 200-301

CCNP Enterprise

CCNP Security

CCIE Enterprise Lab

CCIE Security Lab

CCNP Service Provider

CCNP Data Center

CCNP Collaboration

CCIE DC Lab

Monitoring port traffic on a Cisco switch is essential for diagnosing network slowdowns, preventing packet drops, and planning link capacity upgrades. Whether you need an instant command-line snapshot during a live outage or continuous historical telemetry across hundreds of access and trunk ports, Cisco switches expose bandwidth data through built-in CLI counters, hardware controllers, and SNMP MIB objects. This guide walks through the exact CLI commands to check port utilization, how to interpret Cisco’s load metrics, platform differences between Catalyst and Nexus switches, troubleshooting saturated links, and automating long-term bandwidth tracking.

Checking Interface Bandwidth Utilization via Cisco CLI Commands

When investigating link congestion on a Cisco Catalyst (IOS or IOS-XE) switch, follow this progression of commands to move from a switch-wide overview down to granular per-port throughput rates.

Step 1: View a Switch-Wide Traffic Snapshot with show interfaces summary

To see real-time throughput across every port at once without scrolling through pages of detailed counters, run show interfaces summary:

Switch# show interfaces summary

 *: interface is up
 IHQ: pkts in input hold queue     IQD: pkts dropped from input queue
 OHQ: pkts in output hold queue    OQD: pkts dropped from output queue
 RXBS: rx rate (bits/sec)          RXPS: rx rate (pkts/sec)
 TXBS: tx rate (bits/sec)          TXPS: tx rate (pkts/sec)
 TRTL: throttle count

  Interface              IHQ   IQD  OHQ   OQD  RXBS RXPS  TXBS TXPS TRTL
------------------------------------------------------------------------
* GigabitEthernet1/0/1     0     0    0     0 45200   62 89400  110    0
* GigabitEthernet1/0/2     0     0    0   142 912M 98400  120M 18500    0

Focus on the RXBS (receive bits per second) and TXBS (transmit bits per second) columns to immediately spot heavily loaded links, alongside OQD (output queue drops) to identify ports actively dropping traffic due to saturation.

Step 2: Check Port Utilization Percentages with show controllers utilization

Rather than manually converting bits per second into percentages, run show controllers utilization to view current receive and transmit utilization expressed directly as a percentage (0–100%) per port, along with total switch backplane utilization:

Switch# show controllers utilization
Port       Receive Utilization  Transmit Utilization
Gi1/0/1             1                    1
Gi1/0/2            91                   12
Gi1/0/3             0                    0

Total Ports : 24
Total Ports Receive Bandwidth Percentage  : 4 %
Total Ports Transmit Bandwidth Percentage : 1 %
Average Switch Percentage Utilization     : 2 %

Step 3: Inspect Detailed Single-Port Throughput with show interfaces

Once you identify a busy port, inspect its exact load and moving-average bit rate using show interfaces <interface-id>:

Switch# show interfaces GigabitEthernet1/0/2
GigabitEthernet1/0/2 is up, line protocol is up (connected)
  Hardware is Gigabit Ethernet, address is 0050.56bf.20a1 (bia 0050.56bf.20a1)
  MTU 1500 bytes, BW 1000000 Kbit/sec, DLY 10 usec,
     reliability 255/255, txload 30/255, rxload 232/255
...
  5 minute input rate 912450000 bits/sec, 98412 packets/sec
  5 minute output rate 120110000 bits/sec, 18504 packets/sec

Step 4: Filter Output Across Ports Using CLI Pipe Modifiers

To scan only the status, load fractions, and throughput rates on one or all ports without extra counter noise, combine show interfaces with regular expression pipe filters:

Switch# show interfaces GigabitEthernet1/0/2 | include line protocol|reliability|rate
GigabitEthernet1/0/2 is up, line protocol is up (connected)
     reliability 255/255, txload 30/255, rxload 232/255
  5 minute input rate 912450000 bits/sec, 98412 packets/sec
  5 minute output rate 120110000 bits/sec, 18504 packets/sec

Omitting the interface ID (show interfaces | include line protocol|reliability|rate) outputs this condensed three-line summary for every interface on the switch.


Interpreting txload, rxload, and the 5-Minute Rate

Running show interfaces exposes raw load fractions and rolling averages that can be misleading if you do not understand how Cisco IOS calculates them. The following frequently asked questions explain how to interpret and tune these metrics.

What do txload and rxload mean on a Cisco switch interface?

txload (transmit load) and rxload (receive load) express interface bandwidth utilization as a fraction of 255 over the configured calculation interval. A reading of 255/255 indicates a 100% saturated link. Because Cisco IOS never displays 0/255 on an active link, 1/255 is the baseline minimum reading and indicates that link utilization is anywhere between 0% and 0.39%.

How does a Cisco switch calculate interface bandwidth percentage?

To convert txload or rxload into a standard percentage, divide the numerator by 255 and multiply by 100:
$$\text{Utilization (\%)} = \left(\frac{\text{Load Value}}{255}\right) \times 100$$
For example, rxload 232/255 equals (232 / 255) * 100 = 90.98%. Alternatively, divide the reported input rate or output rate (in bits/sec) by the interface’s BW (bandwidth) parameter. Keep in mind that Cisco calculates txload and rxload against the configured bandwidth <kbps> value on the interface. While physical access ports automatically update BW based on negotiated link speed (such as 1000000 Kbit/sec for 1 Gbps), logical interfaces (SVIs, Port-Channels, Tunnels) or manually overridden ports may calculate inaccurate txload/rxload ratios if the bandwidth statement does not match actual physical capacity.

Why can the default “5 minute input/output rate” hide network congestion?

The 5-minute input and output rates are exponentially weighted moving averages calculated over a 300-second window. Because traffic is smoothed across five minutes, sub-second or multi-second traffic spikes—known as microbursts—are averaged out. A switch port can suffer severe buffer exhaustion and packet loss during brief bursts while still displaying a low 5-minute bit rate and txload 1/255.

How do I change load-interval 30 for more accurate real-time utilization?

To make txload, rxload, and the input/output rate reflect current traffic conditions rather than a 5-minute average, lower the interface load-interval down to its minimum value of 30 seconds (configurable in multiples of 30 seconds):

Switch# configure terminal
Switch(config)# interface GigabitEthernet1/0/2
Switch(config-if)# load-interval 30
Switch(config-if)# end

Once applied, show interfaces switches its output line to 30 second input rate and 30 second output rate, updating load calculations ten times faster to capture short-term traffic surges.


Comparing Interface Utilization Commands: Catalyst IOS-XE vs. Nexus NX-OS

While enterprise Catalyst switches run Cisco IOS or IOS-XE, data center Cisco Nexus switches run NX-OS, which provides enhanced built-in rate tables and live terminal monitoring. The table below compares how port utilization is checked across both operating systems.

Monitoring FeatureCisco Catalyst (IOS / IOS-XE)Cisco Nexus (NX-OS)
Detailed Single-Port Utilizationshow interfaces <id>show interface <id>
Default Averaging WindowSingle 300-second window (5 minute input/output rate)Dual windows displayed simultaneously by default (30 seconds and 5 minutes)
Switch-Wide Rate & Percentage Tableshow interfaces summary (shows bps/pps) and show controllers utilization (shows %)show interface counters rate (combines Rx-Mbps, Rx-%, Tx-Mbps, and Tx-% in one table)
Live Interactive CLI MonitorNot available natively in classic CLI (requires repeated manual execution or Tcl/EEM scripts)monitor interface [ethernet <slot/port>] (refreshes live in a full-screen top-style terminal view)
Load Interval Configurationload-interval <30-600> (supports 1 interval per interface)load-interval counter {1 \| 2 \| 3} <5-300> (supports up to 3 concurrent intervals down to 5 seconds)
Quick CLI Rate Filtershow interfaces \| include line protocol\|rateshow interface \| include Ethernet\|rate

Cisco CLI Command Reference and Counter Field Definitions

For precise verification against Cisco technical specifications, this section defines the exact command syntax, counter fields, and official reference documentation used for interface bandwidth auditing.

Command Syntax Specifications

  • show interfaces [type number] [counters [errors | protocol status | trunk]] [summary] [stats]
    Queries interface status, hardware queue buffers, moving-average bit/packet rates, and load ratios.
  • show interfaces [type number] counters
    Displays cumulative 64-bit inbound and outbound octet counts alongside frame classification totals (unicast, multicast, broadcast) since the last counter reset.

Standard Field Definitions

  • BW: Configured interface bandwidth expressed in kilobits per second (Kbit/sec). Used strictly as the denominator for txload, rxload, EIGRP metric calculation, and SNMP utilization formulas; does not rate-limit physical hardware throughput.
  • reliability: Exponentially weighted 5-minute moving average of frame delivery reliability expressed as an integer fraction of 255 (255/255 = 100% reliable), derived from input errors, CRC/FCS failures, and frame alignment errors relative to total frames received.
  • txload / rxload: Exponentially weighted moving average of outbound and inbound channel utilization expressed as an integer fraction of 255, calculated over the active load-interval.
  • input rate / output rate: Layer 2 throughput rate measured in bits/sec and packets/sec over the active load-interval, inclusive of Ethernet framing overhead.
  • InOctets / OutOctets: Cumulative 64-bit count of all valid Layer 2 bytes (including MAC header, payload, and 4-byte FCS) received or transmitted on the interface.
  • Total output drops: Cumulative number of packets discarded because the outbound hardware egress buffer reached capacity. Rising output drops alongside low average txload confirms microburst congestion.

Official Cisco Documentation References

  • Cisco IOS Interface and Hardware Component Command Reference: Authoritative specification for show interfaces, load-interval, and hardware queue definitions across IOS and IOS-XE releases.
  • Cisco Catalyst 9000 Series Switches Command Reference: Platform-specific syntax reference for show controllers utilization, show interfaces counters, and UADP ASIC buffer counters.
  • Cisco Document ID 8141 (How to Calculate Bandwidth Utilization Using SNMP): Official specification detailing half-duplex and full-duplex utilization equations using MIB-II ifTable and 64-bit ifXTable counters.

Troubleshooting High Interface Utilization and Identifying Top Talkers

Once you confirm that a switch port is saturated, knowing the utilization percentage alone is not enough—you must identify which hosts or applications are consuming the bandwidth. Follow these sequential steps to isolate the root cause.

Step 1: Classify the Traffic Type (Unicast, Multicast, or Broadcast)

Check whether the spike is driven by normal point-to-point data transfers, multicast streams, or a Layer 2 broadcast storm by running show interfaces <id> counters:

Switch# show interfaces GigabitEthernet1/0/2 counters

Port            InOctets    InUcastPkts    InMcastPkts    InBcastPkts
Gi1/0/2      84920144102      612045100         120400          45102

Port           OutOctets   OutUcastPkts   OutMcastPkts   OutBcastPkts
Gi1/0/2      11042100512       89401200          84120          12004

If InBcastPkts or InMcastPkts are climbing rapidly alongside utilization, investigate Layer 2 bridging loops, Spanning Tree topology changes, or unpruned multicast video floods. If InUcastPkts or OutUcastPkts dominate, proceed to host identification.

Step 2: Trace Connected Hosts via MAC and ARP Tables

Identify the source devices connected to the saturated port by querying the MAC address table:

Switch# show mac address-table interface GigabitEthernet1/0/2
          Mac Address Table
-------------------------------------------
Vlan    Mac Address       Type        Ports
----    -----------       ----        -----
  10    0050.569a.4c12    DYNAMIC     Gi1/0/2

If multiple MAC addresses appear on the port, check show cdp neighbors GigabitEthernet1/0/2 detail or show lldp neighbors GigabitEthernet1/0/2 to determine if the port is an uplink to another switch or hypervisor. To resolve the MAC address to an IP address, run show ip arp | include 0050.569a.4c12 on the Layer 3 switch or default gateway (or show device-tracking database interface Gi1/0/2 if IP Device Tracking is active).

Step 3: Query Top Talkers Using Flexible NetFlow or NBAR2

If Flexible NetFlow (FNF) is configured on the interface, inspect the local flow cache directly from the CLI sorted by byte count to view the exact source/destination IP pairs and TCP/UDP ports responsible for the spike:

Switch# show flow monitor MONITOR-IN cache sort highest counter bytes top 10

On Catalyst switches with Application Visibility and Control (AVC / NBAR2) enabled on the port, identify the exact Layer 7 application consuming the bandwidth:

Switch# show ip nbar protocol-discovery interface GigabitEthernet1/0/2 stats bit-rate top-n 10

Step 4: Run an On-Box Embedded Packet Capture (EPC)

If NetFlow is not enabled on a saturated Layer 2 access port, use IOS-XE Embedded Packet Capture (EPC) to sample live headers directly in switch memory without setting up an external Wireshark SPAN port:

Switch# monitor capture CAP interface GigabitEthernet1/0/2 both
Switch# monitor capture CAP match any
Switch# monitor capture CAP limit duration 10
Switch# monitor capture CAP start

After 10 seconds, inspect the captured packets directly in the terminal using show monitor capture CAP buffer brief to see which IP hosts are dominating the link, then clean up with no monitor capture CAP.


Automating Utilization Checks with SNMP OIDs and Python

To poll port utilization programmatically across multiple switches without manual CLI logins, use standard 64-bit SNMP MIB counters or automated CLI scraping scripts.

64-Bit SNMP OIDs and Full-Duplex Calculation Template

Always poll 64-bit high-capacity (ifXTable) counters on Gigabit and faster interfaces, as legacy 32-bit ifInOctets counters wrap around to zero in under 34 seconds on a 1 Gbps link.

# SNMP MIB-II (ifXTable) 64-Bit OIDs (Replace <ifIndex> with port index)
ifName        : 1.3.6.1.2.1.31.1.1.1.1.<ifIndex>
ifHCInOctets  : 1.3.6.1.2.1.31.1.1.1.6.<ifIndex>   (64-bit Inbound Bytes)
ifHCOutOctets : 1.3.6.1.2.1.31.1.1.1.10.<ifIndex>  (64-bit Outbound Bytes)
ifHighSpeed   : 1.3.6.1.2.1.31.1.1.1.15.<ifIndex>  (Interface Speed in Mbps)

# Full-Duplex Bandwidth Utilization Formula (Two polls separated by delta_seconds)
Input_Utilization_Pct  = ((ifHCInOctets_t2 - ifHCInOctets_t1) * 8 * 100) / (delta_seconds * (ifHighSpeed * 1000000))
Output_Utilization_Pct = ((ifHCOutOctets_t2 - ifHCOutOctets_t1) * 8 * 100) / (delta_seconds * (ifHighSpeed * 1000000))

Python Netmiko Script to Export Cisco Switch Port Utilization to CSV

This script connects to a list of Cisco IOS/IOS-XE switches via SSH, parses the interface rate and load metrics covered earlier, and writes a consolidated utilization report to a CSV file.

import csv
import re
from netmiko import ConnectHandler

SWITCHES = [
    {
        "device_type": "cisco_ios",
        "host": "192.168.1.10",
        "username": "admin",
        "password": "YOUR_PASSWORD",
    },
]

INTERFACE_PATTERN = re.compile(
    r"^(?P<intf>\S+) is up, line protocol is up.*?"
    r"BW (?P<bw>\d+) Kbit/sec.*?"
    r"txload (?P<txload>\d+)/255, rxload (?P<rxload>\d+)/255.*?"
    r"input rate (?P<in_bps>\d+) bits/sec.*?"
    r"output rate (?P<out_bps>\d+) bits/sec",
    re.DOTALL | re.MULTILINE,
)

def collect_utilization():
    rows = []
    for device in SWITCHES:
        host = device["host"]
        with ConnectHandler(**device) as conn:
            output = conn.send_command("show interfaces")
            for block in re.split(r"\n(?=\S+ is )", output):
                match = INTERFACE_PATTERN.search(block)
                if match:
                    data = match.groupdict()
                    rx_pct = round((int(data["rxload"]) / 255) * 100, 2)
                    tx_pct = round((int(data["txload"]) / 255) * 100, 2)
                    rows.append([
                        host,
                        data["intf"],
                        int(data["bw"]),
                        int(data["in_bps"]),
                        int(data["out_bps"]),
                        rx_pct,
                        tx_pct,
                    ])

    with open("switch_port_utilization.csv", "w", newline="") as f:
        writer = csv.writer(f)
        writer.writerow([
            "Switch_Host", "Interface", "Bandwidth_Kbps",
            "Input_Rate_bps", "Output_Rate_bps", "RX_Util_Pct", "TX_Util_Pct"
        ])
        writer.writerows(rows)

if __name__ == "__main__":
    collect_utilization()

Best SNMP Network Monitoring Tools for Historical Port Graphing

While CLI commands and ad-hoc scripts are ideal for on-demand checks, dedicated network monitoring platforms poll 64-bit SNMP counters continuously to graph long-term trends and alert on threshold breaches. Here are the top tools for monitoring Cisco switch port utilization:

  1. Paessler PRTG Network Monitor: Best for rapid deployment with minimal configuration. PRTG auto-discovers Cisco switch interfaces using 64-bit SNMP Traffic sensors and offers pre-built Cisco NetFlow/IPFIX sensors to correlate bandwidth spikes with top talkers. It includes a permanently free tier for up to 100 sensors, making it ideal for monitoring critical uplinks in small-to-medium networks.
  2. SolarWinds Network Performance Monitor (NPM): Best for large enterprise Cisco environments. SolarWinds NPM provides deep native integration with Catalyst and Nexus switches, tracking interface utilization, output queue drops, and CBQoS policy metrics alongside automated Layer 2/Layer 3 topology mapping. When paired with SolarWinds NetFlow Traffic Analyzer (NTA), clicking a saturated switch port graph immediately displays the conversations causing the spike.
  3. Zabbix: Best free, open-source enterprise monitoring platform. Zabbix includes official out-of-the-box Cisco IOS and NX-OS SNMP templates that use Low-Level Discovery (LLD) to automatically detect active switch ports, graph inbound/outbound bps using ifHCInOctets/ifHCOutOctets, and trigger alerts when utilization exceeds custom thresholds.
  4. LibreNMS: Best turnkey open-source network monitoring system built specifically for network hardware. LibreNMS uses CDP and LLDP to auto-discover Cisco switch topologies, generates granular RRDtool historical bandwidth graphs per interface without manual template tuning, and includes built-in 95th-percentile bandwidth calculation.
  5. Prometheus (snmp_exporter) + Grafana: Best for high-frequency polling and microburst detection. Whereas legacy SNMP tools commonly poll every 1 to 5 minutes, Prometheus paired with snmp_exporter or Streaming Telemetry (gNMI on Catalyst 9000 and Nexus 9000 switches) can scrape interface counters every 5 to 15 seconds, visualizing short-lived traffic bursts on customizable Grafana dashboards.

Conclusion

Checking interface utilization on a Cisco switch starts with quick CLI visibility using show interfaces summary and show controllers utilization, followed by deeper per-port analysis with show interfaces and a 30-second load-interval to expose short-term bursts. When high bandwidth usage is detected, combining counter classification, MAC/ARP lookups, Flexible NetFlow, or Embedded Packet Capture allows you to pinpoint the exact host and application saturating the link, while 64-bit SNMP polling and monitoring platforms ensure you maintain continuous visibility across your switching infrastructure.

Please follow and like us:
Last modified: November 17, 2026

Author

Comments

Write a Reply or Comment

Your email address will not be published.