What is Grafana? Dashboards explained and your lab

You’ll build two Ubuntu Server VMs. graf01 (192.168.56.70) will run Grafana, Nginx, Prometheus, Loki, InfluxDB and Mailpit as the course goes on, and app01 (192.168.56.71) plays the application server with a web site, MySQL and the agents that feed graf01. Each VM gets a NAT adapter for the internet and a Host-only adapter so the VMs and your laptop reach each other on fixed addresses.

Already did the SaatVedha Prometheus course? Your prom01 and node01 VMs can do the same job, but every command in this course uses graf01 and app01 and their addresses, so building fresh VMs keeps things simpler.

1Check the VirtualBox host-only network

In VirtualBox Manager open File > Tools > Network Manager and pick the Host-only Networks tab. You should see one adapter with IPv4 address 192.168.56.1 and mask 255.255.255.0. Untick its DHCP server, because both VMs get static addresses.

Run this on your laptop (Linux or macOS terminal, or PowerShell on Windows):

VBoxManage list hostonlyifs

You should see:

Name:            vboxnet0
GUID:            786f6276-656e-4074-8000-0a0027000000
DHCP:            Disabled
IPAddress:       192.168.56.1
NetworkMask:     255.255.255.0
IPV6Address:
IPV6NetworkMaskPrefixLength: 0
HardwareAddress: 0a:00:27:00:00:00
MediumType:      Ethernet
Wireless:        No
Status:          Up
VBoxNetworkName: HostInterfaceNetworking-vboxnet0

Your GUID and hardware address will differ. On Windows the name looks like “VirtualBox Host-Only Ethernet Adapter” instead of vboxnet0, and that’s fine. What matters is the 192.168.56.1 address and DHCP Disabled.

✅ Checkpoint: A host-only network with 192.168.56.1/24 exists and its DHCP server is off.

2Create graf01

Click New, name it graf01, pick the Ubuntu Server ISO and tick Skip Unattended Installation so you see the installer. Before the first start, open Settings > Network: Adapter 1 attached to NAT, Adapter 2 enabled and attached to Host-only Adapter on the 192.168.56.1 network.

graf01 gets the bigger share because Grafana, Prometheus and Loki all run there later. Lesson 13 asks you to raise it to 4 GB for a short while, and on an 8 GB laptop that still leaves room for your browser.

✅ Checkpoint: graf01 exists with 3 GB RAM, 2 CPUs and two adapters, NAT and Host-only.

3Install Ubuntu Server on graf01

Start the VM and accept the installer defaults. Use hostname graf01, user student, and tick Install OpenSSH server. Leave the second card (enp0s8) unconfigured, because netplan sets it in the next step, then remove the ISO, reboot and log in.

Run this on graf01:

hostnamectl --static
ip -br addr

You should see:

graf01
lo               UNKNOWN        127.0.0.1/8 ::1/128
enp0s3           UP             10.0.2.15/24 metric 100 fe80::a00:27ff:fe52:8d1e/64
enp0s8           DOWN

enp0s3 is the NAT card with VirtualBox’s usual 10.0.2.15 address and a route to the internet. enp0s8 is the host-only card with no address yet.

✅ Checkpoint: You are logged in as student on graf01 and enp0s8 is listed.

4Give enp0s8 a static address with netplan

Paste this whole block on graf01:

sudo tee /etc/netplan/60-hostonly.yaml >/dev/null <<'EOF'
network:
  version: 2
  ethernets:
    enp0s8:
      dhcp4: false
      addresses: [192.168.56.70/24]
EOF
sudo chmod 600 /etc/netplan/60-hostonly.yaml
sudo netplan apply
ip -br addr show enp0s8

You should see:

enp0s8           UP             192.168.56.70/24 fe80::a00:27ff:fe9b:3c10/64

There’s no gateway on this card on purpose, because internet traffic keeps using the NAT card. The chmod 600 matters too: netplan warns “Permissions for /etc/netplan/60-hostonly.yaml are too open” when other users can read the file.

✅ Checkpoint: enp0s8 on graf01 shows 192.168.56.70/24 and state UP.

5Clone app01 and give it its own identity

Shut graf01 down with sudo poweroff, right-click it and choose Clone. Name the clone app01, select Generate new MAC addresses for all network adapters and a Full clone, then set its RAM to 1536 MB and 1 CPU. Start app01 alone and run:

Run this on app01:

sudo hostnamectl set-hostname app01
sudo sed -i 's/192.168.56.70/192.168.56.71/' /etc/netplan/60-hostonly.yaml
sudo netplan apply
sudo rm -f /etc/machine-id && sudo systemd-machine-id-setup
sudo rm -f /etc/ssh/ssh_host_* && sudo dpkg-reconfigure openssh-server
hostnamectl --static; ip -br addr show enp0s8

First you’ll see “Initializing machine ID from random generator.” and then a “Creating SSH2 RSA key; this may take some time …” line followed by its fingerprint, and the same for the ECDSA and ED25519 keys. The last two lines are the ones to check:

app01
enp0s8           UP             192.168.56.71/24 fe80::a00:27ff:fe1d:72e4/64

A clone carries the same machine ID and SSH host keys as the original. Making new ones stops duplicate DHCP leases on the NAT side and the “host key changed” warning when you SSH in later. Disk size stays 30 GB on the clone, which is more than app01 needs and harmless.

✅ Checkpoint: app01 has hostname app01, address 192.168.56.71 and a fresh machine ID.

6Add both names to /etc/hosts and test

Start graf01 again. On both VMs add the two lines below, then ping across by name.

Run this on graf01 and on app01:

printf '192.168.56.70 graf01 graf01.lab\n192.168.56.71 app01 app01.lab\n' | sudo tee -a /etc/hosts
ping -c 2 app01

On graf01 the output looks like this (your ping times will differ):

192.168.56.70 graf01 graf01.lab
192.168.56.71 app01 app01.lab
PING app01 (192.168.56.71) 56(84) bytes of data.
64 bytes from app01 (192.168.56.71): icmp_seq=1 ttl=64 time=0.587 ms
64 bytes from app01 (192.168.56.71): icmp_seq=2 ttl=64 time=0.463 ms

--- app01 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.463/0.525/0.587/0.062 ms

From your laptop, ssh student@192.168.56.70 and ssh student@192.168.56.71 should both work now. Copy and paste is much easier in an SSH window than in the VirtualBox console, so use SSH from here on.

✅ Checkpoint: graf01 and app01 ping each other by name and you can SSH into both from your laptop.

7Take the base snapshot

Shut both VMs down (sudo poweroff), select each one in VirtualBox Manager and use Snapshots > Take. Name both snapshots base, so any later lesson can roll back to a clean, networked Ubuntu in a few seconds.

Or run this on your laptop:

VBoxManage snapshot graf01 take base --description "Ubuntu + netplan + hosts"
VBoxManage snapshot app01 take base --description "Ubuntu + netplan + hosts"
VBoxManage snapshot graf01 list

You should see (your UUIDs will differ):

0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...100%
Snapshot taken. UUID: 3f0c2a51-7d2e-4b8e-9a6e-2c1f5d7b9e40
0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...100%
Snapshot taken. UUID: 8b1d6e93-2a4c-4f07-b5d8-61e0c9a7f312
   Name: base (UUID: 3f0c2a51-7d2e-4b8e-9a6e-2c1f5d7b9e40) *
   Description: Ubuntu + netplan + hosts

The star marks the current snapshot. Start both VMs again before the next step.

✅ Checkpoint: Both VMs have a snapshot called base and are running again.

8A first look at Grafana from the tarball (temporary)

On graf01, download the Grafana OSS 13.3.0 standalone build, unpack it in your home folder and start it in the foreground. Lesson 2 installs Grafana properly from the APT repository with a service user and systemd, so this copy is only a test drive and you delete it afterwards.

Run this on graf01:

cd ~
wget -q https://dl.grafana.com/grafana/release/13.3.0/grafana_13.3.0_40412481033_linux_amd64.tar.gz
tar -zxf grafana_13.3.0_40412481033_linux_amd64.tar.gz
cd grafana*/ && ls -d bin conf public

You should see:

bin  conf  public

The tarball unpacks into one folder named after Grafana and its version, and bin, conf and public are the three parts Grafana needs to start. The long number in the file name is a build ID that changes with each release, so copy the current link from grafana.com/grafana/download with OSS selected if this one has moved on. The commands use grafana*/ so you don’t have to type the folder name.

Now start it (still on graf01, inside that folder):

./bin/grafana server

The output is long. These are the lines to look for (your times, commit and build date will differ):

logger=settings t=2026-11-08T06:10:42.512734918Z level=info msg="Starting Grafana" version=13.3.0 commit=... branch=HEAD compiled=...
logger=sqlstore t=2026-11-08T06:10:42.530118274Z level=info msg="Created default admin" user=admin
logger=http.server t=2026-11-08T06:10:45.981402117Z level=info msg="HTTP Server Listen" address=[::]:3000 protocol=http subUrl= socket=

Between them you’ll also see “Config loaded from”, “Path Home” and “Path Data” lines that point into the folder you unpacked.

Grafana logs in logfmt: logger= names the part of Grafana that wrote the line, then the time, level and message. The line you wait for is “HTTP Server Listen” address=[::]:3000, which means it accepts connections on port 3000 on all addresses. Without a grafana.ini it uses conf/defaults.ini and keeps its SQLite database under data/ in that folder.

✅ Checkpoint: The log shows “HTTP Server Listen” with address [::]:3000 and Grafana keeps running.

9Log in and draw a graph from TestData

On your laptop browse to http://192.168.56.70:3000. Log in as admin with password admin; Grafana then asks for a new password, and for this throwaway copy you can set one or click Skip. Ubuntu’s firewall (ufw) is off on a fresh install, which is why the port is reachable at all.

Open Connections > Data sources > Add new data source, search for TestData, select it and click Save & test. TestData is a built-in source that makes up fake series, so you can learn panels before any real backend exists.

Now open Explore from the left menu, choose TestData at the top, leave the scenario on Random Walk and click Run query. A wobbly line appears for the last hour, and changing the time picker to “Last 6 hours” redraws it.

Back on graf01, in a second SSH session, run this:

curl -s http://localhost:3000/api/health; echo
ls ~/grafana*/data | grep grafana.db

You should see:

{
  "database": "ok",
  "version": "13.3.0",
  "commit": "..."
}
grafana.db

The “commit” value is the Git hash of the build, so yours shows a real hash there. The health endpoint needs no login and is what load balancers and scripts check. The grafana.db file now holds your admin user and the TestData data source, and nothing else, because the random walk numbers are made up on each query and never stored.

✅ Checkpoint: Explore shows a Random Walk line from TestData, and /api/health returns “database”: “ok”.

10Stop and clean up the test copy

Go back to the first SSH session and press Ctrl+C. Grafana logs a “Shutdown started” line and exits. Then remove the test copy so it doesn’t get in the way of the real install in Lesson 2.

Run this on graf01:

cd ~
rm -rf ~/grafana*/ ~/grafana_13.3.0_*_linux_amd64.tar.gz
sudo ss -ltnp | grep ':3000' || echo "port 3000 is free"

You should see:

port 3000 is free

✅ Checkpoint: Nothing listens on port 3000 and the tarball folder is gone from your home directory.