Colin Holyoake and Graham Eames at IBM Hursley first implemented IBM Turbonomic in November of 2021. Within just a couple of hours of the installation, 6,000 virtual systems started to appear in the team’s instance of Turbonomic. The team finally had a singular view into their full data center application stack from the application (virtual machine/container) all the way through to the associated storage. Among other entities outside x86 virtualization in Hursley’s environment, Turbonomic monitors IBM FlashSystem, and Red Hat OpenShift running on IBM z15.
“IBM Turbonomic is almost like a sixth sense,” explains Holyoake, Certified Datacenter Design and Sustainability Manager at IBM Hursley. “It gives you a full visibility of your entire environment and how it’s running. It doesn’t compromise the resiliency at all. It assesses, advises, and improves performance.”
With Turbonomic, the IBM Hursley team has the ability to quickly identify inter-dependencies of the logical environment throughout the physical data center and reallocate resources to assure performance. Within 24 hours of installing Turbonomic, for example, the team identified a critical performance risk and resizing opportunity. They observed that in one of their hosting clusters each node containing 60 cores and 2 TB RAM was experiencing excessive CPU ready states and was negatively impacting development operations. Using Turbonomic’s resourcing recommendations, the IBM Hursley team then identified lightly used resources elsewhere in their environment which could be reallocated to reduce this performance risk and they implemented those recommendations right away.
An important component of this full-stack visibility is the range of integrations IBM Turbonomic offers. For example, the IBM Hursley team implemented the Instana® integration as well as the Red Hat® OpenShift® integration. A large percentage of the workloads on that cluster with critical performance risks consisted of Red Hat OpenShift-based workloads. Deploying Turbonomic at this level of the stack provided much deeper visibility of those workloads and helped the team quickly resolve the performance issue. This level of visibility has also helped the team tune the workload sizings within OpenShift via automation and better balance performance demands with their available resources on an ongoing basis.