Install Node Exporter on Amazon Linux 2023: Prometheus Monitoring for Tableau Server

The fastest way to get system metrics from a Tableau Server EC2 instance into Prometheus is Node Exporter running as a systemd service, discovered automatically via ec2_sd_configs. No hardcoded IPs, no manual config updates when instances cycle.

This guide walks through every step on Amazon Linux 2023: downloading the binary, creating a locked-down system user, writing the systemd unit, and wiring up Prometheus with EC2 service discovery so the instance appears as UP in your targets page.

---

Prerequisites

• EC2 instance running Amazon Linux 2023 with root or sudo access
• Prometheus server with network access to port 9100 on the target instance
• EC2 instance tags (Name, environment) set on the Tableau Server instance
• IAM permissions for Prometheus to call ec2:DescribeInstances

---

Step 1: Download Node Exporter

cd /tmp
curl -LO https://github.com/prometheus/node_exporter/releases/download/v1.10.2/node_exporter-1.10.2.linux-amd64.tar.gz

Always grab the latest release from the official Node Exporter releases page. At the time of writing, v1.10.2 (released October 25, 2025) is the latest stable version.

---

Step 2: Extract and Install the Binary

tar xzf node_exporter-1.10.2.linux-amd64.tar.gz
mv node_exporter-1.10.2.linux-amd64/node_exporter /usr/local/bin/
chmod +x /usr/local/bin/node_exporter

Verify the binary is in place:

which node_exporter
# /usr/local/bin/node_exporter

ls -la /usr/local/bin/node_exporter
# -rwxr-xr-x. 1 root root 22919216 Oct 25 2025 /usr/local/bin/node_exporter

---

Step 3: Create a Dedicated System User

Node Exporter should never run as root. Create a locked-down, non-login user:

useradd --no-create-home --shell /bin/false node_exporter

Skipping this step causes the service to fail with status=217/USER when systemd tries to start it.

---

Step 4: Create the systemd Service File

cat > /etc/systemd/system/node_exporter.service <<'EOF'
[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 \
  --collector.filesystem.mount-points-exclude='^/(sys|proc|dev|host|etc)($$|/)' \
  --web.listen-address=:9100

Restart=on-failure
RestartSec=5s

[Install]
WantedBy=multi-user.target
EOF

Key flags explained:

Flag	Purpose
--collector.filesystem.mount-points-exclude	Excludes virtual filesystems like /proc, /sys, /dev from disk metrics
--web.listen-address=:9100	Default port Prometheus scrapes from
Restart=on-failure	Auto-restarts if the process crashes

---

Step 5: Enable and Start the Service

systemctl daemon-reload
systemctl enable --now node_exporter

--now starts the service immediately and enable ensures it starts on every reboot automatically.

---

Step 6: Verify It Is Running

systemctl status node_exporter

Expected output:

● node_exporter.service - Node Exporter
     Loaded: loaded (/etc/systemd/system/node_exporter.service; enabled; preset: disabled)
     Active: active (running) since Thu 2026-06-19 10:00:00 UTC; 5s ago
   Main PID: 461 (node_exporter)
      Tasks: 5 (limit: 19660)
     Memory: 2.9M

Then confirm the metrics endpoint is reachable:

curl -s http://localhost:9100/metrics | head -20

You should see lines starting with # HELP and # TYPE followed by metric values.

---

Step 7: Useful systemctl Commands

# Start
systemctl start node_exporter

# Stop
systemctl stop node_exporter

# Restart
systemctl restart node_exporter

# Status
systemctl status node_exporter

# Follow logs live
journalctl -u node_exporter -f

# Last 100 log lines
journalctl -u node_exporter -n 100

---

Step 8: Configure Prometheus with ec2_sd

Using ec2_sd_configs means Prometheus discovers the instance automatically from its EC2 tags - no hardcoded IPs that break when instances are replaced.

Add this job to your prometheus.yml:

scrape_configs:
  - job_name: 'node-exporter'
    ec2_sd_configs:
      - region: us-east-1  # replace with your AWS region
        port: 9100
    relabel_configs:
      # Only scrape instances with this Name tag
      - source_labels: [__meta_ec2_tag_Name]
        regex: tableau-server
        action: keep
      # Use the Name tag as the instance label
      - source_labels: [__meta_ec2_tag_Name]
        target_label: instance
      # Carry the environment tag through
      - source_labels: [__meta_ec2_tag_environment]
        target_label: env

Then reload Prometheus without a full restart:

systemctl reload prometheus

Navigate to Status > Targets in the Prometheus UI. The Tableau Server instance should appear as UP.

---

Step 9: Verify EC2 Tags

The ec2_sd relabeling depends entirely on your instance having the right tags. Check them:

aws ec2 describe-instances \
  --filters "Name=tag:Name,Values=tableau-server" \
  --query "Reservations[].Instances[].Tags" \
  --output table

The output should include at minimum:

Tag Key	Value
Name	tableau-server
environment	production

If the tags are missing or misspelled, the instance will not match the keep relabel rule and will not appear in Prometheus targets.

---

Troubleshooting

Error	Cause	Fix
Unit node_exporter.service not found	Service file missing	Create /etc/systemd/system/node_exporter.service then run daemon-reload
status=217/USER	node_exporter OS user does not exist	Run useradd --no-create-home --shell /bin/false node_exporter
Active: activating (auto-restart)	Binary or user missing	Check ls /usr/local/bin/node_exporter and that the user exists
Prometheus target shows DOWN	Port 9100 blocked	Open TCP 9100 inbound in the EC2 security group from your Prometheus server IP
No targets in Prometheus UI	EC2 tags missing or wrong	Verify tags with aws ec2 describe-instances

---

Security Checklist

• Node Exporter runs as a non-root dedicated user (node_exporter)
• Port 9100 should only be open to your Prometheus server IP in the security group - not 0.0.0.0/0
• This setup does not configure TLS, which is fine for internal VPC traffic. If you are exposing Node Exporter externally, add --web.config.file with TLS settings

---

What to Monitor on Tableau Server

Once Node Exporter is scraped, these are the metrics that matter most for a Tableau Server host:

Metric	What It Tells You
node_cpu_seconds_total	CPU saturation during extract refreshes and heavy queries
node_memory_MemAvailable_bytes	Memory pressure - Tableau is memory-hungry
node_filesystem_avail_bytes	Disk space for workbooks, extracts, and temp files
node_disk_io_time_seconds_total	Disk I/O bottlenecks during extract processing
node_network_receive_bytes_total	Network throughput for data source connections

Pair these with Tableau Server’s own process metrics (via the Tableau REST API or the built-in tabadmin status) for full visibility into both the OS layer and application layer.

---

Tested on: Amazon Linux 2023, kernel 6.1.161, Node Exporter v1.10.2

---

Get Full Observability and Support for Your Tableau Environment

System metrics are one layer. A production Tableau setup also needs performance tuning, workbook optimization, extract scheduling, and someone who knows the platform inside out.

Our Tableau professional services cover everything from initial deployment to ongoing managed support:

• Monitoring and alerting - Node Exporter, Prometheus, and Grafana dashboards tailored to Tableau Server health
• Performance tuning - Extract refresh optimization, memory pressure analysis, and query performance
• Managed operations - Upgrades, backups, user management, and incident response

We also provide Prometheus consulting and Grafana consulting if you need the full observability stack built around your Tableau environment.

Talk to our Tableau team →