Proxmox monitoring is the practice of tracking the health of a Proxmox VE server or cluster and the virtual machines and containers running on it. It covers node-level resources such as CPU, memory, and storage pool usage, cluster state such as quorum and node membership, operational outcomes such as backup job results and task errors, and the resource consumption of each individual guest.

• Cluster quorum loss: A cluster that drops below quorum stops making configuration changes and cannot start high availability workloads.
• Silent backup failures: A backup job that stops completing leaves you without a restore point, and nothing in the Proxmox interface interrupts you to say so.
• Storage pool exhaustion: A full storage pool halts snapshots, backups, and guest writes across every VM that depends on it.
• Noisy neighbor guests: One VM consuming CPU or disk I/O degrades every other guest on the same node.
• Task and log errors: Failed tasks and repeated system log errors mark problems that have not yet become outages.

The Proxmox VE Wizard is available in Nagios XI 2026R1.3 and later. It connects to the Proxmox API and creates Host and Service objects from the selections you make in the wizard.

Proxmox Server and Cluster Monitoring

• CPU Usage: Node CPU usage as a percentage, checked against Warning and Critical thresholds you set.
• Memory Usage: Node memory usage as a percentage.
• Storage Usage: Storage pool usage on the Proxmox server.
• Cluster Health: Cluster quorum and node status. The check returns Critical when quorum is lost.
• Backup Status: Backup job status and age.
• Log Errors: A count of errors and warnings in the Proxmox system logs over a lookback window.
• Task Errors: A count of failed Proxmox tasks over a lookback window.

Proxmox VM and Container Monitoring

Each selected VM or container becomes its own Host object in Nagios XI rather than a Service on the Proxmox server Host, so guests carry their own status, history, notifications, and graphs.

• CPU Usage: Guest CPU usage as a percentage.
• Memory Usage: Guest memory usage as a percentage.
• Disk I/O: Read, write, and total disk activity for the guest, reported as data volume.
• Network I/O: Inbound, outbound, and total network traffic, reported cumulatively since the guest started.
• Storage Volume: Disk space usage inside the guest. Pair it with hard drive monitoring on the underlying nodes to see both layers.
• Load Average: Node load average at one, five, and fifteen minutes.
• Uptime: Guest uptime, which exposes unplanned restarts.
• Backup Status: Backup job status and age for the individual guest.

Documentation: How To Monitor Proxmox VE With Nagios XI

Proxmox Thresholds and Alerting

Every metric you enable adds a threshold section to the wizard where you set the Warning and Critical values for that check. Guest CPU defaults to 80 percent Warning and 95 percent Critical. The Log Errors check defaults to five errors Warning and ten Critical over a five-minute lookback, and the Task Errors check defaults to one error Warning and three Critical over the same window. Lookback times accept 1 to 240 minutes.

Set the Log Errors lookback conservatively. A longer lookback window increases the load the check places on the Proxmox server.

Proxmox Authentication and Access

Authenticate with a Proxmox API Token in the format user@realm!tokenid=secret, or with a Proxmox username and password. An API Token scoped to monitoring is the better choice, because it can be revoked without touching a user account. The Privilege Separation setting on the token in Proxmox governs which metrics the token can read, so review those privileges first when a token returns partial data.

The wizard also offers a Verify SSL Certificate option for the connection to the Proxmox API.

Guests and hosts in one place. Proxmox nodes, virtual machines, and containers land in the same Nagios XI instance as your physical servers, network devices, databases, and applications, so an incident is investigated in one interface instead of three.

Per-guest history. Because each VM and container is a Host object, you get its own performance graphs and state history. That turns “the VM feels slow” into a chart showing when its CPU usage crossed the threshold.

Backups you can prove ran. Backup Status checks on the server and on individual guests surface a job that stopped completing, rather than leaving it to be discovered during a restore.

Alerting you control. Thresholds, check intervals, notification schedules, and escalations are configured per Service, so the team that owns the cluster hears about quorum loss and the team that owns an application hears about its guest.

Self-hosted. Nagios XI runs on your own hardware, inside your own network, with Proxmox API credentials that never leave it.

• Nagios XI version: 2026R1.3 or later, which is where the Proxmox VE Wizard is available.
• Python library: The proxmoxer library installed on the Nagios XI server.
• Network path: Access from the Nagios XI server to the Proxmox server on TCP port 8006, or on the custom port your Proxmox installation uses.
• Credentials: A Proxmox API Token, or a Proxmox username and password. Store the value as <API_TOKEN> in any configuration management you keep outside Nagios XI.
• Token privileges: Privilege Separation configured on the token so it can read the metrics you intend to check.
• Guest addressing: A valid IP address or hostname for each VM you select. QEMU/KVM guests populate this automatically. Guests that do not populate it need it entered in the wizard, otherwise the ping check that determines whether the Host is Up will not run.

1. Install the Python prerequisites. On the Nagios XI server, as root, run pip3 install proxmoxer. On Debian 12 and Ubuntu 24, run pip3 install proxmoxer --break-system-packages. On a RHEL 8 RPM or offline installation, confirm Python 3.9 is installed, symlink python3 to it, then install urllib3, requests, and proxmoxer with python3.9 -m pip.
2. Open the wizard. Navigate to Configure > Configuration Wizards, search for proxmox, and select the Proxmox VE wizard.
3. Enter connection details. In Step 1, supply the Proxmox server hostname or IP address, the port if it differs from 8006, and either an API Token or a username and password. Choose whether to verify the SSL certificate, then click Test Connection and confirm the OK response before continuing.
4. Select metrics and guests. In Step 2, set a Host Name for the Proxmox server, choose the server and cluster metrics you want, choose the VM and container metrics you want, set the Warning and Critical thresholds that appear for each, then select the VMs and containers to monitor. Re-running the wizard against an existing Host Name appends checks to that Host.
5. Finish the common steps. Review the summary at the top of Step 3, define check intervals, notification settings, and groups across Steps 3 to 5, then click Finish to apply the configuration.

Documentation: How To Monitor Proxmox VE With Nagios XI | Understanding and Using Configuration Wizards in Nagios XI

The Proxmox VE Wizard reads state and metrics from the Proxmox API and compares each result against the thresholds you set. It reports on the hypervisor layer and on guest resource consumption. To monitor what is running inside a guest, such as services, processes, and application state, install the Nagios Cross-Platform Agent (NCPA) on the guest and add checks through Linux server monitoring or the equivalent wizard for its operating system.

The Log Errors check counts errors in the Proxmox system logs over a short window. Collecting, indexing, searching, and retaining those logs is the job of Nagios Log Server, which is licensed separately. Traffic analysis across the virtual and physical network, using NetFlow, sFlow, J-Flow, or IPFIX, belongs to Nagios Network Analyzer.

Infrastructure Layers

• Server Monitoring
• Linux Monitoring
• Hard Drive Monitoring

Protocols and Applications

• SNMP Monitoring
• SQL Server Monitoring

Run the Proxmox VE Wizard against your own cluster and see what it finds. The Nagios XI trial is fully functional for 30 days with no Node limit, and a Nagios engineer is available to you for the length of it. For licensing questions, email [email protected].

• Start Your 30-Day Nagios XI Trial
• Request a Personalized Demo
• View Pricing and Plans