How To Install Prometheus on Ubuntu 26.04 LTS

If your server crashes at 3 a.m. and you find out from an angry client instead of an alert, you have a monitoring problem. Most sysadmins learn this lesson the hard way, usually right before a deadline. That is exactly why Prometheus exists, and why so many Linux teams are moving toward it for real time visibility into CPU, memory, disk, and network health.

This guide walks you through how to install Prometheus on Ubuntu 26.04 using the official binaries, a dedicated system user, and a proper systemd service. You will also learn how to set up Prometheus on Ubuntu 26.04 for host monitoring by adding Node Exporter, opening the right firewall ports, and confirming that everything is actually working, not just running.

By the end, you will have a working Prometheus instance that starts on boot, scrapes metrics correctly, and survives a reboot without falling over. We will also cover common mistakes people make when they try to configure Prometheus on Ubuntu 26.04, so you can skip the frustrating trial and error phase entirely.

This is written as a practical Linux server tutorial, not a theory lecture. Every command has a reason behind it, because copying commands without understanding them is how servers break in production.

Table of Contents

Prerequisites

Before you start, make sure you have the basics covered. Skipping these steps is the number one reason installations fail halfway through.

A server or VM running Ubuntu 26.04 LTS, with at least 2 CPU cores and 4 GB of RAM for light monitoring workloads.
Root or sudo access to the machine, since Prometheus needs system level directories and a dedicated service user.
A stable internet connection to download the Prometheus binary release.
Basic familiarity with the terminal, including navigating directories and editing files.
curl or wget installed for downloading files.
tar installed for extracting archives, which ships with Ubuntu by default.
An open port plan, since Prometheus listens on port 9090 and Node Exporter listens on port 9100.

If you are working on a fresh Ubuntu 26.04 install, you likely already meet these requirements. If not, we will fix any gaps in Step 1.

Step 1: Update Your System

Before installing anything new, you need a clean base to work from. An outdated package list is one of the most common causes of weird dependency errors later.

Refresh package information

Run the following command to update your local package index:

sudo apt update

This command does not install anything yet. It just tells your system where to find the latest versions of available packages.

Apply available upgrades

Next, apply any pending updates:

sudo apt upgrade -y

Why this matters: Prometheus interacts closely with the kernel and system libraries through exporters. Running on an outdated system increases the risk of compatibility issues, especially with newer monitoring tools that expect current library versions.

If the kernel gets updated during this step, a reboot is a good idea before continuing:

sudo reboot

Wait a minute, then reconnect using SSH if you are working remotely.

Step 2: Create a Dedicated Prometheus User

This step gets skipped a lot, and it should not be. Running monitoring software as root is a bad habit that creates unnecessary security risk.

Create the system user

Run this command to create a user with no login shell:

sudo useradd --no-create-home --shell /usr/sbin/nologin prometheus

Why this matters: Prometheus does not need to log in interactively. Giving it a restricted, no-shell account limits what an attacker could do if the service were ever compromised. This is a basic security practice that senior sysadmins apply to almost every background service.

Create required directories

Prometheus needs two main locations, one for configuration and one for stored metrics data:

sudo mkdir /etc/prometheus
sudo mkdir /var/lib/prometheus

/etc/prometheus will hold configuration files, alert rules, and console templates.
/var/lib/prometheus will store the actual time series database files.

Separating these two locations makes backups and troubleshooting much easier down the road. If your data directory grows too large, you can manage it without touching your configuration files.

Step 3: Download and Install Prometheus Binaries

Now for the main event. Instead of using outdated third party repositories, we will pull the official release directly from the source.

Check the latest version

Visit the official Prometheus GitHub releases page in your browser to confirm the current stable version number. As of mid 2026, the actively supported release lines are Prometheus 3.5 and newer 3.x versions, so always check before downloading.

Download the archive

Once you know the version number, download it with curl. Replace the version number below with the current one you found:

curl -LO https://github.com/prometheus/prometheus/releases/download/v3.5.0/prometheus-3.5.0.linux-amd64.tar.gz

Why this matters: Downloading directly from the official GitHub releases page avoids third party mirrors that may be outdated or, in rare cases, tampered with. This keeps your monitoring stack trustworthy from the start.

Extract the files

Unpack the archive using tar:

tar xvf prometheus-3.5.0.linux-amd64.tar.gz

You should see output listing each extracted file, including the prometheus binary, the promtool binary, and default configuration folders.

Move the binaries into place

Copy the two main executables into a standard system path:

sudo mv prometheus-3.5.0.linux-amd64/prometheus /usr/local/bin/
sudo mv prometheus-3.5.0.linux-amd64/promtool /usr/local/bin/

Why this matters: Placing binaries in /usr/local/bin keeps them separate from software managed by the Ubuntu package manager. This avoids conflicts if you later install a different Prometheus version through apt.

Move configuration files

Move the sample config and support directories into /etc/prometheus:

sudo mv prometheus-3.5.0.linux-amd64/consoles /etc/prometheus
sudo mv prometheus-3.5.0.linux-amd64/console_libraries /etc/prometheus
sudo mv prometheus-3.5.0.linux-amd64/prometheus.yml /etc/prometheus

Set correct ownership

Now assign ownership of these directories to the Prometheus user you created earlier:

sudo chown -R prometheus:prometheus /etc/prometheus
sudo chown -R prometheus:prometheus /var/lib/prometheus

Skipping this step is a common reason the service fails to start. If Prometheus cannot read or write to its own directories, it will exit immediately.

Step 4: Configure Prometheus

This is where most tutorials get lazy. A good configuration file is the difference between a monitoring system that actually helps you and one that just sits there collecting dust.

Open the configuration file

Edit the file with your preferred text editor:

sudo nano /etc/prometheus/prometheus.yml

Review the default structure

The file should contain a basic structure like this:

global:
  scrape_interval: 15s
  evaluation_interval: 15s

scrape_configs:
  - job_name: "prometheus"
    static_configs:
      - targets: ["localhost:9090"]

Understand each setting

scrape_interval controls how often Prometheus collects metrics from each target.
evaluation_interval controls how often alert rules are checked against the collected data.
job_name is a label used to identify a group of monitored targets.
static_configs and targets define which addresses Prometheus should scrape.

Why this matters: A shorter scrape interval gives more detailed data but uses more disk space and CPU. A longer interval saves resources but may miss short lived spikes. Fifteen seconds is a solid starting point for most small to medium servers.

Save the file and exit the editor.

Step 5: Create a systemd Service for Prometheus

Running Prometheus manually from the terminal works for testing, but it will not survive a reboot or an SSH disconnect. A systemd service fixes that.

Create the service file

sudo nano /etc/systemd/system/prometheus.service

Add the service definition

Paste the following configuration:

[Unit]
Description=Prometheus Monitoring System
Wants=network-online.target
After=network-online.target

[Service]
User=prometheus
Group=prometheus
Type=simple
ExecStart=/usr/local/bin/prometheus \
  --config.file=/etc/prometheus/prometheus.yml \
  --storage.tsdb.path=/var/lib/prometheus/ \
  --web.console.templates=/etc/prometheus/consoles \
  --web.console.libraries=/etc/prometheus/console_libraries

[Install]
WantedBy=multi-user.target

Why this matters: The ExecStart line tells systemd exactly how to launch Prometheus, including where to find its config file and where to store its data. Without this, systemd would not know how to run the program correctly, and it would fail on boot.

Save the file and exit.

Reload systemd

Tell systemd to recognize the new service file:

sudo systemctl daemon-reload

Enable and start the service

sudo systemctl enable prometheus
sudo systemctl start prometheus

Verify the service status

sudo systemctl status prometheus

You should see output showing active (running) in green text, along with the process ID and recent log lines. If you see failed instead, check the troubleshooting section below.

Step 6: Open Firewall Access and Verify in the Browser

Your service is running, but you still need to confirm it is reachable and actually collecting data.

Allow the Prometheus port

If you are using UFW, allow port 9090:

sudo ufw allow 9090/tcp

Why this matters: Prometheus will not be reachable from outside the server if the firewall blocks its port. Only open this port to trusted networks in production, since the default setup has no built in authentication.

Access the web interface

Open a browser and go to:

http://your-server-ip:9090

You should see the Prometheus web interface load successfully.

Check the Targets page

Click on Status, then Targets. You should see the prometheus job listed with a state of UP in green.

Why this matters: A green UP status confirms that Prometheus is successfully scraping metrics from itself. If it says DOWN, something is wrong with your config file or the service is not fully started.

Step 7: Install Node Exporter for Host Metrics

Prometheus alone only monitors itself. To track real server health like CPU load, memory usage, and disk space, you need Node Exporter.

Download Node Exporter

curl -LO https://github.com/prometheus/node_exporter/releases/download/v1.8.2/node_exporter-1.8.2.linux-amd64.tar.gz

Extract and install

tar xvf node_exporter-1.8.2.linux-amd64.tar.gz
sudo mv node_exporter-1.8.2.linux-amd64/node_exporter /usr/local/bin/

Create a dedicated user

sudo useradd --no-create-home --shell /usr/sbin/nologin node_exporter

Create a systemd service

sudo nano /etc/systemd/system/node_exporter.service

Paste this configuration:

[Unit]
Description=Node Exporter
Wants=network-online.target
After=network-online.target

[Service]
User=node_exporter
Group=node_exporter
Type=simple
ExecStart=/usr/local/bin/node_exporter

[Install]
WantedBy=multi-user.target

Then enable and start it:

sudo systemctl daemon-reload
sudo systemctl enable node_exporter
sudo systemctl start node_exporter

Add Node Exporter as a scrape target

Go back into your Prometheus config file:

sudo nano /etc/prometheus/prometheus.yml

Add a new job under scrape_configs:

  - job_name: "node_exporter"
    static_configs:
      - targets: ["localhost:9100"]

Why this matters: Without this addition, Prometheus has no idea Node Exporter exists. Adding the job tells Prometheus where to pull server level metrics from, which is the whole point of setting this up.

Save the file, then restart Prometheus so it picks up the change:

sudo systemctl restart prometheus

Refresh the Targets page in your browser. You should now see both prometheus and node_exporter listed as UP.

Troubleshooting Common Issues

Even careful setups run into snags. Here are the most common problems and how to fix them.

Service fails to start with a permission error
This usually means the Prometheus user does not own its own data or config directories. Re-run the chown commands from Step 3 and try starting the service again.
Config file errors on startup
Run promtool check config /etc/prometheus/prometheus.yml before restarting the service. This catches YAML formatting mistakes before they cause a silent failure.
Web interface will not load in the browser
Check that the firewall port is open and that the service is actually running with sudo systemctl status prometheus. Also confirm you are using the correct server IP address, not localhost, if accessing remotely.
Target shows DOWN on the Targets page
This usually means the exporter service is not running, or the port number in your config does not match the exporter’s actual listening port. Double check with sudo systemctl status node_exporter.
Disk space fills up quickly
Prometheus stores all scraped data locally by default. If storage grows too fast, consider adding a retention flag like --storage.tsdb.retention.time=15d to the systemd service file to automatically delete old data.