Monitoring virtual machines with Prometheus vs. Graphite

Monitoring virtual machines with Prometheus vs. Graphite

Table of Contents

  1. Introduction
  2. What are virtual machines?
    1. Virtualization types
  3. Why is monitoring virtual machines important?
  4. Monitoring metrics for virtual machines
    1. Availability
    2. CPU usage
    3. Memory utilization
    4. Top virtual machines
  5. Benefits of Prometheus and Hosted Graphite
    1. Prometheus
    2. Graphite
    3. Hosted Prometheus and Graphite
  6. Wrapping up


Virtual machines give you a flexible and convenient environment where people can access different operating systems, networks, and storages while still using the same computer. This prevents them from purchasing extra machines, switching to other devices, and maintaining them. This helps companies to save costs and increase task efficiency.


Although using VMs for everyday tasks may be enjoyable, ensuring consistent performance and performing maintenance can be daunting. The key to success for smooth operation is monitoring. Monitoring gives you the visibility you need to diagnose the current status and lets you know what actions need to be taken. 


This article discusses the effective ways of monitoring virtual machines with Prometheus and Graphite. MetricFire offers a comprehensive monitoring solution with Hosted Graphite where you can gain in-depth insight into your VM resources with minimal configuration. 


If you would like to learn more about it, please book a demo with us, or sign up for a free trial today.


What are virtual machines?

The first step of setting up a good monitoring architecture is understanding your target systems or software. In our case, it is virtual machines. A virtual machine (VM) can consist of various resources such as CPU, memory, storage, operating systems, and hardware. These elements in VM don’t exist in a physical form but they are defined as software to mimic a physical environment.  


Virtual machines are widely used by enterprises and individual users. The virtual machine products below are some of the most popular ones in the market. You may want to check each to find what’s suitable for your use case.


  • VMware Workstation Player
  • VirtualBox
  • Parallels Desktop
  • QEMU
  • Citrix Hypervisor


Virtualization types

In virtualization, there are five main types. When you understand the differences between them, you can design and focus on the monitoring metrics that matter the most.


  • Hardware virtualization: hardware virtualization means you virtualize computers and operating systems and consolidate them into a single physical server. A hypervisor in that server allows those VMs to share the resources and to run multiple VMs simultaneously. Hardware virtualization is also known as server virtualization.
  • Software virtualization: you can virtualize operating systems so that a physical computer with hardware can switch to a different OS without having to leave the computer. For example, on your Mac computer, you can use Windows.
  • Storage virtualization: you can also virtualize several physical storage devices to appear as a single storage device. With this, you can achieve enhanced performance, faster speed, load balancing, and cost reduction.
  • Network virtualization: when you have multiple sub-networks, you can virtualize them into a single software-based virtual network resource. Network virtualization can divide bandwidth into multiple isolated channels. Each of them can be assigned to servers and devices.
  • Desktop virtualization: this type of virtualization separates a desktop environment from a physical machine so that the virtualized desktop environment can be used anywhere on any device. As people gain easy accessibility to the desktop, benefits like cost savings and easy management can be attainable.



Why is monitoring virtual machines important?

Virtual machine monitoring is the activity of monitoring the individual VMs and VMs created in a virtualization environment. Monitoring helps your VM resources to run optimally. As you monitor resource status, you can know what’s happening in your resources and avoid serious incidents by planning ahead. 


As we monitor virtual machines, we can gain benefits in the three major areas.


  • Resource planning: by monitoring, administrators can plan capacity better. The planning includes checking if the current resource is appropriate for your VMs and how much more will be needed in the future. If admins find much idle resources, they can re-allocate them to where it’s needed, which is also cost-efficient.
  • Overload prevention: when you run multiple virtual machine software in a single server, the resources of the server can easily be depleted slowing down all the VMs. Monitoring can help you to control resource use and allocate some of the VMs into other servers if needed.
  • Stable performance: Virtualization monitoring stabilizes the performance of virtual machines. It allows you to identify key problems through various metrics.


As businesses monitor its VM resources, companies can operate their IT resources efficiently and optimally.


Monitoring metrics for virtual machines

Although there are discrete types of virtualization and virtualization software, there are monitoring metrics that are commonly important. By looking through each metrics, you can have an idea on how you want to design your monitoring stack.



Availability monitoring helps you to prevent adverse situations by checking the uptime of virtual machines and notifying operators of issues before they start to have a negative impact on users and business. 


CPU usage

CPU utilization is an important aspect in monitoring. The usage status shows how much computation each virtual machine performs. There can be busy times when CPU usage spikes and idle times when virtual machines don’t do computation. Using these metrics, you can allocate resources more effectively and decide whether or not to increase CPU capacity.


Memory utilization

When your virtual machines run memory-intensive jobs, it can cause out-of-memory. When it is a virtualized server environment, overusing memory (CPU too) can lead to slowness in all the virtual machines sharing the resources in the server. You may want to set up notifications to identify high memory usage and abnormally low use of memory.   


Top virtual machines

If you want more focused metrics, displaying top virtual machines can improve your productivity. Let’s say you want to see the top VMs that are consuming the server CPU the most. Displaying the top CPU-consuming virtual machines can point you to where to focus and where not to. You can take this approach for top users of memory, disk I/O, and network.


Benefits of Prometheus and Hosted Graphite

We learned there are important monitoring metrics for VMs. The next important task is setting up your monitoring tool. For this, Prometheus and Graphite are recommended choices for this task. Let’s first find out what Prometheus and Graphite are.



This open-source monitoring tool was born in 2012. Since then, many organizations have started using it to monitor their infrastructures. Prometheus has an active developer and user community that provides updates and supports. Prometheus is able to collect and store metrics as time series data. This means it can store data in key-value pairs with the timestamp when the pairs were captured. Among many features, Prometheus provides the following.


  • You can incorporate a multi-dimensional data model with time series data that can be identified by metric name and key-value pairs.
  • Prometheus has PromQL that’s a flexible query language to utilize the multi-dimensionality.
  • It does not have to rely on distributed storage. Prometheus can have autonomous single server nodes.
  • Prometheus collects time-series data by pulling over HTTP.
  • You can also push time series data through an intermediary gateway.
  • Targets can be discovered by service discovery or static configuration.
  • Various graphs and dashboards are supported.



Graphite is also an open-source monitoring software that is designed to perform reliably on affordable hardware or Cloud infrastructure. People use Graphite to track the performance of their resources such as websites, applications, network servers, and more. This tool was developed by Chris Davis in 2006 and was released under Apache 2.0 license in 2008. 


Among many strengths, Graphite is known to provide a scalable solution for real-time graphing. You can write an application that digests numeric time-series data and sends it to Graphite’s processing backend - carbon that stores the data in its database. Then, the data can be visualized through Graphite’s web interfaces.


Hosted Prometheus and Graphite

Prometheus and Graphite both offer a good range of monitoring features and allow you to track performances of your resource in real-time. This real-time monitoring capability is an important factor for monitoring virtual machines since you can have the most up-to-date data. In spite of the benefits, they inherently have a common difficulty. That is maintenance due to the nature of open-source projects. When you use an open-source software, starting from installation to implementation, all become your tasks. It will become an ongoing cost and require continuous efforts to keep them running smoothly.


If your objective is to monitor your VM resources, using Hosted Graphite will help you to focus on what matters the most by reducing maintenance efforts. MetricFire robust provides hosting services for Graphite software so that users can forget about the logistical burden and pay more attention to the monitoring job itself.


Wrapping up

In this article, we covered what virtual machines are and learned the important monitoring metrics. Prometheus or Graphite can help you to achieve your monitoring goal and run VM resources smoothly. Although they both have powerful monitoring and metrics features, it can be demanding to start setting up and to maintain them by yourself. To bring more efficiency to your team, try our Hosted Graphite service. 


Visit MetricFire today and check out our Hosted Graphite service.

You might also like other posts...
metricfire May 28, 2023 · 9 min read

How To Monitor A Linux Virtual Machine

You can ensure your Linux virtual machine is operating at its peak efficiency and... Continue Reading

metricfire May 24, 2023 · 11 min read

Best Grafana dashboard for Graphite Metrics

See examples of displaying Graphite metrics in Grafana dashboards. Use Hosted Graphite and Hosted... Continue Reading

metricfire May 22, 2023 · 7 min read

Exploring cAdvisor for Common Use Cases

cAdvisor can help you monitor your container deployments. But what is cAdvisor exactly, and... Continue Reading

header image

We strive for
99.999% uptime

Because our system is your system.

14-day trial 14-day trial
No Credit Card Required No Credit Card Required