Introduction
In the first 2 parts of this blog series we discussed the importance of measurement and only provisioning the required resources in reducing consumption. This blog looks at what you can do to reduce consumption once already deployed and, although focused on the features of the Nutanix Cloud Manager™ (NCM) product, it may help give ideas to IT practitioners using other technology platforms.
Operational Tools for Improving Efficiency
NCM is part of the Nutanix Cloud Platform (NCP) technology stack that helps IT teams better manage and report on their infrastructure with features that operate across environments running on Nutanix Cloud Infrastructure, VMware virtualization and Public Cloud. A unified solution for providing intelligent operations, self-service and orchestration, security compliance and visibility, and control of cloud costs, but we’ll be looking at a specific subset of those features that can help better manager the consumption of resources and plan for the future consumption of resources, ultimately helping consumers of IT “take only what they need”.
Categories
A great place to start is to use Categories to facilitate VM centric operations. Although this is not going to directly produce any reduction in resource utilization on its own, using categories is the first step to setting up policies to manage various aspects of your Nutanix Cloud Platform more efficiently.
For example, by setting replication policies per VM to meet the required RPO, but not excessively exceed it, you can ensure you’re only replicating what is necessary and not wasting resources (including network bandwidth) replicating more data than you need to.
Categories are part of standard Prism Central (NCM not required) but can also be used as the basis for more easily taking advantage of many of the features of NCM including reporting.
Right Sizing
Use the efficiency reports in Nutanix Cloud Manager to help ensure that provisioning is optimal and that you are not wasting resources. These reports can easily be configured to be automatically emailed to consumers / department heads so that they can easily take and/or approve appropriate action and keep your IT resources streamlined and performance.
Although resizing a single VM on its own will unlikely have much of a measurable effect on reducing the overall consumption of hardware resources, the cumulative effect of keeping many VMs optimally provisioned can reduce overall cluster utilization significantly as well as make future hardware capacity planning exercises more accurate. It’s also a great way to keep your Procurement Department colleagues happy, with the ability to demonstrate that an efficient operating model is being used, and that they’re only being asked to approve an expansion once it really is required!
X-Play Automation
Use Nutanix X-Play (“Crossplay”) automation to then automatically implement changes out of hours with a minimum of manual effort. Or completely automate the detection and resolution of a potential VM capacity issue.
X-Play can also be used to easily turn VMs on or off, upgrade their resources or automate other processes. Turning off idle VMs perhaps only reduces power consumption incrementally, but if many VMs can be powered off, then this can add up significantly. Always be mindful of whether the outcome is worth the effort. Turning VMs off for a short period of time might not actually represent a net benefit as both power up and shutdown processes can result in a utilization spike. As with any change, it’s important to measure the result as well as estimate the intended benefit ahead of implementation. Reviewing results should help improve future estimates that can be used to help prioritize future initiatives.
Resource Planning
Use capacity runway in Nutanix Cloud Manager, including its “What If” scenario planning, to add in future workloads. A lot of IT departments try to only purchase new hardware when they refresh their infrastructure platforms every 3-5 years, often going through lengthy, manual planning and purchasing processes to ensure that they won’t run out of capacity. This is becoming increasingly hard to do, especially in economic climates where mergers, acquisitions or divestitures are frequent. By operationalizing resource planning with a tool like capacity runway you can change your approach to capacity management and infrastructure purchasing by keeping on top of your resource management more regularly and growing more granularly. Regular reports can be scheduled or they can be automated should a cluster approach the limits of its available resources.
Operationalized capacity management should be a key part of any sustainable infrastructure management strategy as it not only provides an opportunity to reduce the energy used to power your IT systems, but also could reduce emissions by helping avoid the purchase and powering of new hardware.
FinOps
Financial Operations is a management practice that shares responsibility for the cost of IT infrastructure across the consumers and users of an organization. By effectively implementing FinOps, organizations can reduce the demand for IT services by making sure that individual departments and business units know what they are consuming and how much it is costing them.
The NCM Cost Governance™ solution is a FinOps tool that offers Chargeback / Showback and cost governance features across public, private and hybrid cloud environments. It can also be used by IT, finance and other business teams to allocate costs across business units at a granular level to help them ensure they are only consuming what they need to. Multi cloud chargeback / showback features of NCM™ are available with a free trial so there’s no reason to not get started!
Anomaly Detection
Use the anomaly detection feature of Nutanix Cloud Manager to detect patterns of behavior or help to analyze periods of high or erratic utilization. E.g. Using CPU performance charts and analyzing in relation to other metrics might help you to understand when part of an operating system, agent or applications are over consuming. A good example of this is where AntiVirus or backup software might generate a peak in CPU usage if not properly configured. Or if a particular database query is causing excessive disk I/O.
Anomaly detection is just one part of the NCM behavioral learning tools that use machine-learning capabilities to analyze resource usage over time and help IT teams improve their management of resources.
Neither NCM or Prism® currently include power consumption as a visible metric in the GUI (measurement strategies are available), but any reduction in resources obviously would have an effect, albeit possibly a marginal one on the overall utilization of a system. Nutanix is actively working to understand how we can better support our customers to more sustainably manage their IT systems.
Workload Attribution
Increasingly customers want to attribute the power utilization / energy consumption of a platform to the individual workloads (VMs, Containers, Storage, etc) for FinOps or GreenOps purposes.
Tying a specific workload VM to a consumed kWh figure is a complex and nuanced process with many different approaches and methodologies. One of the advantages of using the Nutanix Cloud Platform is that nearly all performance and utilization metrics are available through the same interface / API endpoint so customers can more easily understand their performance and capacity profile.
For more advanced users, some of these features can be accessed programmatically via the Prism Central APIs,allowing data analysis to be combined with application or other business data:
Capacity Runway API: https://www.nutanix.dev/api_references/prism-central-v3/#/1b233f90bb7b6-get-scenarios-name-and-uuid-with-pagination
General Performance Stats: https://developers.nutanix.com/api-reference?namespace=aiops&version=v4.r0.a1 (this is in early access, but there are other ways to obtain this data)
These are just some of the related APIs that you could use to get data into your business analysis tools to really understand and optimize your platform consumption. More information on Nutanix’s APIs and examples of how to use them can be found at https://www.nutanix.dev/
Conclusion
On their own, individual changes are unlikely to move the needle if you are looking to reduce the consumption of IT resources in a measurable way. But, if some of these strategies can be operationalised and implemented holistically, making them part of day-to-day IT activities, then over time they can make a significant impact. The key is having both the right tools and the resources to implement them effectively. If the tools are simple to deploy and provide easy to interpret metrics then successful operationalization of the processes needed to make these kinds of changes is much more likely.
Additionally, reducing consumption, especially if it means avoiding purchasing new hardware, almost always means a reduction in costs. By making a start sooner rather than later, you and your colleagues will be quicker on a path to producing a meaningful return on the time invested.