Coping With the Real Challenges of Monitoring Virtualized Infrastructure
Virtualization can make monitoring infrastructure challenging. Besides the fact that you still need to monitor the physical servers and storage that underlie your virtual servers and networks, virtualized environments are far more dynamic than physical infrastructure by design. Virtual machines (VMs) may be brought online, shut down, and moved between physical servers unpredictably. Hypervisors provide features to track VM status, but may not scale effectively when virtual infrastructure grows.
Understand Your Environment and Identify What Needs to Be Monitored
Before investigating tools to support monitoring your virtual infrastructure, you need to decide what needs to be monitored. This is partly a question of identifying the infrastructure components: physical servers, virtual machines, storage devices, and applications; and partly a question of identifying the performance metrics that need to be monitored.
Some tools can focus on monitoring current metrics, such as CPU, memory, disk, and network utilization at the current moment in time. Other tools capture the history of performance metrics over time, allowing trends to be reviewed and potential capacity issues to be addressed before they noticeably impact the performance end users experience.
Choose Your Monitoring Tools
Once you’ve decided what you need to monitor, you can start looking for tools to do this most effectively. You may decide that the hypervisor’s tools, which can monitor the up/down status of a VM and restart it upon failure, is all you need. You may be able to integrate tools you’ve been using to monitor physical devices so they can monitor your VMs as well, or you may wish to obtain new monitoring tools specifically designed for VMs. Financial considerations regarding the costs of licenses as well as the technical capabilities of these tools need to be considered.
Because server monitoring software often provides server management capabilities as well, software designed specifically for virtual machines may provide additional capabilities you find useful, such as automatically deploying VMs without the need to manually configure the hardware or moving VM workloads to optimize the physical server utilization. Hypervisors typically include these functions, but may not work as effectively at scale as a third-party product. Hypervisors also know only about the handful of VMs they’re managing and a third-party product may give a better overall view.
Don’t Ignore the Physical World
Monitoring the virtual machines needs to be integrated with monitoring the physical infrastructure in order to provide a complete picture of utilization and investigate and resolve performance problems.
VM failures can be due to power failures, memory errors, operating system errors, and other problems on the underlying physical server. Performance issues in an application running on a VM may be due to another VM on the same physical box using too much CPU, bandwidth issues on the network, or other causes. Having as much data as possible presented and correlated through a single tool will make identifying and correcting problems simpler and faster.
Get the Benefits of Effective VM Monitoring
With an effective monitoring strategy that keeps an eye on both your physical and virtual infrastructure, you can maximize your VM’s performance and uptime, as well as collect data to help you make smart capacity management decisions. With both VMWare Certified and Microsoft Certified IT Professional (MCITP)-Virtualization members on the Prescient Solutions team, we’re expert at building, monitoring, and managing virtualized environments. Contact us to talk about solving your virtual machine monitoring and management challenges.