Linux7 min readPublished: 9 May 2024Updated: 9 Nov 2026

nload: Monitor Live Network Traffic on Linux

Install nload and read live inbound and outbound bandwidth on a Linux server, plus when iftop, nethogs or vnstat answer the question nload cannot.

G7Cloud Engineering
Platform team
Share:
Abstract illustration representing nload: monitor live network traffic on linux.
  • nload draws live incoming and outgoing bandwidth for each network interface as a scrolling graph in the terminal.
  • Install it with sudo apt install nload, or sudo dnf install nload after enabling EPEL.
  • Press the left and right arrow keys to switch interface. Press q to quit.
  • nload shows how much traffic there is, never which process caused it. For that you need nethogs or iftop.
  • It keeps no history. For last month's usage you need vnstat, installed before you need it.

What is nload and what does it show?

nload is a small terminal tool that shows live network traffic for one interface (a network connection such as eth0) at a time. Incoming traffic fills the top half of the screen and outgoing the bottom, updating about twice a second, with current, average, minimum, maximum and total figures.

What nload looks like
Device eth0 [203.0.113.10] (1/2):
=========================================================
Incoming:
                        Curr: 2.14 MBit/s
     .:||:.             Avg: 1.87 MBit/s
  .:||||||:.            Min: 0.31 MBit/s
 .::||||||::.           Max: 8.92 MBit/s
:::||||||||:::          Ttl: 4.21 GByte
---------------------------------------------------------
Outgoing:
                        Curr: 18.60 MBit/s
      .:|||:.           Avg: 14.22 MBit/s
   .:|||||||:.          Min: 1.05 MBit/s
 .::|||||||||::.        Max: 62.10 MBit/s
::::|||||||||::::       Ttl: 31.80 GByte
Outgoing far higher than incoming is normal for a web server: small requests in, whole pages out.
FieldMeaningWhat to watch for
CurrTraffic right nowA steady figure near your port speed means the link is full
AvgAverage since nload startedThe number to quote, not Curr
MinLowest reading seenA floor well above zero means constant background traffic
MaxHighest reading seenBrief peaks are normal and rarely a problem
TtlTotal moved since nload startedResets to zero every time you restart nload

Ttl counts only from when nload started, so it is not a total for the day.

nload answers how much, never who

It reads traffic counters from the kernel for each interface. It can tell you the link is busy, but nothing about which process, connection or visitor is responsible. When the question is who, skip to the comparison further down.

How do I install nload?

Install
# Debian / Ubuntu
sudo apt update && sudo apt install -y nload

# AlmaLinux / Rocky / CentOS Stream (nload lives in EPEL)
sudo dnf install -y epel-release
sudo dnf install -y nload

# Alpine
sudo apk add nload

# Confirm
nload --version
On the RHEL family you need the EPEL repository first, because nload is not in the base repositories. See the apt and dnf guide if adding a repository is new to you.

It reads counters, so it needs no special rights

nload reads /proc/net/dev, which any user can read, so it does not need sudo. iftop and nethogs do need root, because they capture the traffic itself rather than reading totals.

How do I use nload?

Run nload with no arguments and it picks the first active interface. On a server with several interfaces, name the one you care about, or the numbers will describe traffic you are not interested in.

Everyday usage
# Default interface
nload

# A specific interface
nload eth0

# Find out what your interfaces are called first
ip -brief link show

# Show in megabits rather than the default
nload -u m eth0

# Refresh every 100ms for a finer graph
nload -t 100 eth0

# Watch several interfaces, arrow keys to switch
nload eth0 eth1 wg0
Interface names vary: eth0, ens3, enp1s0, and wg0 for a WireGuard private network, are all common. ip -brief link show is the quickest way to check.
Key or flagDoes
Left / right arrowSwitch to the previous or next interface
q or Ctrl+CQuit
F2Options screen
-u mReport in MBit/s (h for human-readable, K, M, G also accepted)
-t 200Refresh interval in milliseconds
-mMultiple interfaces on one screen, no graphs

Use -m for a quick look at every interface at once.

Pro Tip

Start nload inside tmux before a large transfer, then detach. Reconnect later and you can see the peak and average the transfer reached, instead of guessing from how long it took.

When should I use iftop, nethogs or vnstat instead?

nload tells you the link is busy. The next question is always which connection or which process, and nload cannot answer either. Three other tools share that job.

ToolAnswersNeeds rootKeeps history
nloadHow much traffic on this interface, right nowNoNo
iftopWhich remote addresses the traffic is going toYesNo
nethogsWhich process is using the bandwidthYesNo
vnstatHow much was used yesterday, last week, last monthNoYes

nethogs is the one people wish they had known about sooner.

The other three
# Which remote hosts, with a running bandwidth bar per connection
sudo apt install -y iftop
sudo iftop -i eth0 -nNP

# Which process. This is usually the answer you want.
sudo apt install -y nethogs
sudo nethogs eth0

# Historical usage. Install it BEFORE you need it: it only
# records from the moment the daemon starts.
sudo apt install -y vnstat
sudo systemctl enable --now vnstat
vnstat -d      # by day
vnstat -m      # by month
vnstat -l      # live
iftop flags: -n skips DNS lookups, -N skips port name lookups, -P shows ports. Without them the display is slow and less useful.

Install vnstat now

It records traffic only from the moment its background service starts. Installing it in the middle of an incident tells you nothing about what came before. It costs almost nothing to run, so put it on every server now and it will have the answer when you need it.

What causes unexpected bandwidth on a web server?

When outgoing traffic is higher than your visitor numbers explain, the cause is usually one of these five. Work down the list.

  1. 1Crawlers and scrapers. Aggressive bots, AI training crawlers and content scrapers can outnumber human visitors several times over on a small site. They pull full pages and images, and analytics that rely on JavaScript never see them.
  2. 2Backups leaving the server. A nightly offsite copy of a 40 GB site is 40 GB of outgoing traffic on a schedule. Compare the timing with your backup window before you assume an attack.
  3. 3Large uncompressed images. One 8 MB banner image served ten thousand times is 80 GB. Check whether images are resized and compressed at all.
  4. 4A hotlinked file. Another site embedding your image or PDF directly means you pay for their traffic. The referrer in the access log names them.
  5. 5A compromised server. A server sending large volumes to unfamiliar destinations, especially at night, may be relaying spam or serving files that are not yours. sudo iftop -nNP shows the destinations straight away.
Finding the cause in the access log
# Top 20 user agents by request count: bots surface fast here
awk -F'"' '{print $6}' /var/log/nginx/access.log \
  | sort | uniq -c | sort -rn | head -20

# Top 20 client IPs
awk '{print $1}' /var/log/nginx/access.log \
  | sort | uniq -c | sort -rn | head -20

# Biggest responses by bytes sent (field 10 in the combined format)
awk '{bytes[$7] += $10} END {for (u in bytes) print bytes[u], u}' \
  /var/log/nginx/access.log | sort -rn | head -20
The third command is the useful one. It ranks URLs by total bytes sent rather than by hits, and bytes are what fill the link.

Stopping bots before they reach your site is cheaper than serving them

Once you know which crawlers are responsible, stop them reaching the application at all. On G7Cloud, ScaleShield checks traffic that looks automated before it reaches your site, so scrapers and bad bots do not tie up your PHP workers. See how ScaleShield works.

If the traffic is genuine and the link is simply full, measure its ceiling with a terminal speed test before you buy more bandwidth. Steady load near the port speed is a capacity problem; brief peaks are not.

Frequently asked questions

What is nload used for?

Watching live incoming and outgoing bandwidth on a Linux server's network interfaces. It draws a scrolling graph in the terminal with current, average, minimum and maximum figures, which helps you see whether a link is full during a transfer or a traffic spike.

How do I install nload on Ubuntu?

Run sudo apt update && sudo apt install -y nload. It is in the standard repositories on Debian and Ubuntu. On AlmaLinux, Rocky or CentOS Stream you need EPEL first: sudo dnf install -y epel-release, then sudo dnf install -y nload.

Can nload show which process is using my bandwidth?

No. nload reads per-interface totals from the kernel, so it knows how much but not where from. Use nethogs for a per-process breakdown or iftop for a per-connection one. Both need root, because they capture the traffic itself.

How do I switch between interfaces in nload?

Press the left and right arrow keys. The header shows which interface you are on and its position, such as (1/2). To start on a specific one, name it: nload eth0. Run ip -brief link show first if you are not sure what your interfaces are called.

Does nload keep a history of bandwidth usage?

No. Its totals reset every time it starts, so it cannot tell you what happened yesterday. Use vnstat for that: it runs in the background and reports by hour, day and month, but only from the moment you install and start it.

Put this into practice on G7Cloud

Every site runs in its own dedicated container behind ScaleShield, with daily backups that are restore-tested every night. Start on the free plan, no card needed.

About G7Cloud Engineering

Platform team

Articles written by the engineers who build and run G7Cloud: UK managed hosting and the AI Website Builder. We write about what we operate every day: containers, backups, databases, and the small-business websites that run on them.

More about G7Cloud →