Proxmox VE Monitoring
Monitor Proxmox VE clusters, nodes, virtual machines, containers, storage, and Ceph health through the Proxmox REST API.
NetCrunch provides integrated monitoring for Proxmox Virtual Environment infrastructure. It collects health, status, and performance data from Proxmox clusters, physical Proxmox nodes, QEMU virtual machines, and LXC containers using the Proxmox REST API.
Proxmox monitoring gives you a centralized and hierarchical view of the infrastructure:
- Cluster status and quorum
- Proxmox node availability and resource usage
- Datastore and storage usage
- Ceph cluster state
- Virtual machine and container performance
- CPU, memory, disk, network, swap, and pressure stall metrics
NetCrunch can connect directly to a single Proxmox node or retrieve information from the whole Proxmox cluster.
Requirements
Proxmox monitoring requires credentials that allow read-only access to the Proxmox VE API.
Supported authentication methods:
- API Token, recommended
- Username and password
The monitoring user requires the PVEAuditor role assigned at the root level /.
This grants read-only access to all cluster resources and is required for complete cluster, node, storage, VM, and container visibility.
This requirement applies to all supported Proxmox user types:
- Local Proxmox users, for example
user@pve - Linux PAM users, for example
user@pam - LDAP or Active Directory users
Use an API token whenever possible. It is safer than storing an interactive user password and is easier to rotate without changing the user's login credentials.
Configuration
NetCrunch includes preconfigured automatic monitoring packages for Proxmox VE. When the node operating system monitor is set to Proxmox VE, NetCrunch can automatically apply the required monitoring settings.
For new nodes, Proxmox VE can be detected automatically. You can also enable Proxmox VE monitoring manually in the node monitoring settings.
After valid Proxmox credentials are provided, NetCrunch detects the Proxmox cluster and prompts you to add it for monitoring.
Cluster monitoring options
- Create atlas nodes for Proxmox nodes
- Automatically adds Proxmox cluster nodes to the Network Atlas for monitoring.
Node monitoring options
- Create nodes for guests
- Automatically adds Proxmox guest systems to the Network Atlas. This applies only to guests with a known IP address.
For QEMU virtual machines, the Proxmox Guest Agent must be running for the guest IP address to be available to NetCrunch.
Cluster Monitoring
Cluster monitoring tracks the health and availability of the Proxmox cluster as a whole.
NetCrunch automatically collects storage usage metrics for the Proxmox cluster and monitors key cluster-level conditions.
Cluster alerts
NetCrunch can alert on the following Proxmox cluster conditions:
- Ceph Cluster Warning
- Ceph Cluster Critical
- Cluster has no Quorum
- Low free space on Datastore
- Any Proxmox node goes offline
Cluster monitoring scope
Cluster monitoring is focused on shared infrastructure state, including:
- Cluster quorum
- Node membership
- Datastore usage
- Cluster-level storage health
- Ceph health, if Ceph is used by the Proxmox environment
Proxmox Node Monitoring
A Proxmox node is monitored by enabling the Proxmox VE OS monitor on the node.
Node monitoring tracks the physical or virtual host running Proxmox VE. It collects system-level metrics and monitors host resource pressure.
Node alerts
NetCrunch can alert on the following Proxmox node conditions:
- High CPU Usage
- High memory usage
- Low free disk space
- High swap usage
- Possible storage bottleneck, based on IO Wait
- Node CPU Pressure Warning, when
cpu > 70%orloadavg / maxcpu > 0.7 - Node CPU Saturation Critical, when
cpu > 90%andloadavg / maxcpu > 1.0
Performance Metrics - Proxmox Node
The Proxmox node sensor collects comprehensive metrics about node resource usage and performance.
Processor metrics
- Processor/% CPU Usage
- Processor/CPU Count
Memory metrics
- Memory/% Used
- Memory/% Free
- Memory/Total Bytes
- Memory/Used Bytes
- Memory/Free Bytes
Disk metrics
- Disk/% Used
- Disk/% Free
- Disk/Total Bytes
- Disk/Used Bytes
- Disk/Free Bytes
System metrics
- System/Uptime
- System/% IO Wait
CPU Load metrics
- CPU Load/1 Minute Average
- CPU Load/5 Minutes Average
- CPU Load/15 Minutes Average
Swap metrics
- Swap/% Used
- Swap/% Free
- Swap/Total Bytes
- Swap/Used Bytes
- Swap/Free Bytes
Virtual Machine and Container Monitoring
The Proxmox/VM Sensor is added automatically when NetCrunch discovers that a node is a Proxmox virtual machine or container.
No additional sensor configuration is required.
Supported guest types:
- QEMU virtual machines
- LXC containers
For QEMU virtual machines, the Proxmox Guest Agent must be installed and running if you want NetCrunch to receive the guest IP address from Proxmox.
Guest alerts
NetCrunch can alert on the following guest conditions:
- High guest processor utilization
- High guest memory usage
- Low free disk space on guest
- CPU fully saturated, all tasks waiting
- Applications waiting for CPU
- Processes waiting for memory
- System stalled due to memory pressure
- Applications waiting for disk
- All operations blocked by disk I/O
Performance Metrics - Virtual Machines and Containers
The Proxmox/VM Sensor collects detailed metrics about guest system resource usage.
Guest Processor metrics
- Guest Processor/% CPU Usage
- Guest Processor/CPU Count
Guest Memory metrics
- Guest Memory/% Used
- Guest Memory/% Free
- Guest Memory/Total Bytes
- Guest Memory/Used Bytes
- Guest Memory/Free Bytes
Guest Swap metrics
Guest swap metrics are available for LXC containers only.
- Guest Swap/% Used
- Guest Swap/% Free
- Guest Swap/Total Bytes
- Guest Swap/Used Bytes
- Guest Swap/Free Bytes
Guest Disk metrics
- Guest Proxmox.Guest.Disk/Read Bytes
- Guest Proxmox.Guest.Disk/Write Bytes
- Guest Proxmox.Guest.Disk/Read Bytes/sec
- Guest Proxmox.Guest.Disk/Write Bytes/sec
Guest Network metrics
- Guest Network/In Bytes
- Guest Network/Out Bytes
- Guest Network/In Bytes/sec
- Guest Network/Out Bytes/sec
Guest CPU Pressure Stall metrics
- Guest CPU Pressure Stall/% Some
- Guest CPU Pressure Stall/% Full
Guest Memory Pressure Stall metrics
- Guest Memory Pressure Stall/% Some
- Guest Memory Pressure Stall/% Full
Guest IO Pressure Stall metrics
- Guest IO Pressure Stall/% Some
- Guest IO Pressure Stall/% Full
Guest System metrics
- Guest System/Process ID
Pressure Stall Metrics
Proxmox guest monitoring includes pressure stall metrics where available.
Pressure stall metrics help identify resource contention that may not be obvious from simple utilization values alone.
- CPU pressure
- Shows how much time tasks are delayed because CPU resources are unavailable.
- Memory pressure
- Shows how much time processes are delayed because memory is unavailable or reclaim activity is blocking execution.
- IO pressure
- Shows how much time tasks are blocked by disk or storage I/O.
The % Some metric indicates that at least some tasks were waiting for the resource.
The % Full metric indicates that all non-idle tasks were waiting, which is usually a more severe condition.
Recommended Monitoring Workflow
Use this sequence when adding Proxmox VE monitoring:
Add or discover Proxmox node
Add the Proxmox node to the Network Atlas or let NetCrunch discover it automatically.
Enable Proxmox VE OS monitoring
Set the OS monitoring type to Proxmox VE if it was not detected automatically.
Provide Proxmox credentials
Use an API token or username and password with the PVEAuditor role assigned at /.
Add the detected cluster
After credentials are verified, NetCrunch detects the Proxmox cluster and prompts you to add it for monitoring.
Enable automatic node creation
Use Create atlas nodes for Proxmox nodes to add Proxmox cluster nodes automatically.
Enable guest discovery
Use Create nodes for guests to add virtual machines and containers to the Atlas automatically when their IP addresses are known.
Review alerts and monitoring packs
Check the automatically assigned monitoring packages and adjust thresholds if your Proxmox environment has unusual resource usage patterns.
Best Practices
- Use API tokens
- Prefer API tokens instead of username and password authentication.
- Assign minimum required permissions
- Use the PVEAuditor role at
/for read-only monitoring access. - Monitor both cluster and nodes
- Cluster monitoring detects shared infrastructure problems, while node monitoring detects host-level resource issues.
- Enable Guest Agent for QEMU VMs
- Without the Proxmox Guest Agent, the IP address of a QEMU virtual machine may not be available to NetCrunch.
- Review datastore thresholds
- Storage alerts should reflect your operational policy. Small Proxmox environments may require different free-space thresholds than large clusters.
- Watch IO Wait and pressure stall metrics
- High CPU usage alone does not explain every performance problem. IO Wait and pressure stall metrics help identify storage and scheduling bottlenecks.
- Add guests to the Atlas when useful
- Guest nodes provide better visibility, but only add them automatically when IP addressing and naming are reliable.
Troubleshooting
Cluster is not detected
Check the following:
- The Proxmox API is reachable from the NetCrunch Server or Probe
- The credentials are valid
- The user or token has the PVEAuditor role assigned at
/ - The selected Proxmox node is a member of the expected cluster
Proxmox nodes are not added automatically
Check whether Create atlas nodes for Proxmox nodes is enabled.
Also verify that NetCrunch can resolve and reach the Proxmox node addresses returned by the Proxmox API.
Guest systems are not added automatically
Check whether Create nodes for guests is enabled.
For QEMU virtual machines, verify that:
- Proxmox Guest Agent is installed
- Proxmox Guest Agent is running
- Proxmox reports the guest IP address through the API
For LXC containers, verify that Proxmox can report the container network address.
Guest metrics are incomplete
Some metrics depend on guest type and Proxmox data availability.
For example:
- Guest swap metrics are available only for LXC containers
- Guest IP detection for QEMU VMs requires the Guest Agent
- Pressure stall metrics depend on guest and host support
Summary
NetCrunch Proxmox monitoring provides hierarchical visibility across the full Proxmox VE stack.
It monitors:
- Proxmox clusters
- Proxmox nodes
- Datastores and storage usage
- Ceph health
- QEMU virtual machines
- LXC containers
- Guest resource usage and pressure stall conditions
By combining cluster-level monitoring, node-level monitoring, and automatic guest discovery, NetCrunch gives administrators a complete operational view of Proxmox infrastructure from one place.