Resource recommendations for Kubernetes

The amount of resources that you need to run Turbonomic is related to the complexity and type of entities that it manages. Entities include objects such as virtual machines, containers, pods, or cloud accounts. Turbonomic takes advantages of containers to scale the platform accordingly.

The following table provides a guideline to the cluster resources that you need to support the entities that you expect to manage in a single Turbonomic instance.

Size	Entity scale	Minimum size per node	Number of nodes	Estimated total memory consumption (GB)	Estimated total CPU consumption (Cores)
Small	Up to 1,000 VMs 5,000 to 80,000 entities	2 CPUs * 16 GB	At least 3	8 to 42	3 to 6
Medium	1,000 to 16,000 VMs 80,000 to 200,000 entities	3 CPUs * 32 GB	At least 3	32 to 91	4 to 9
Large	16,000 to 40,000 VMs 200,000 to 1.5 million entities	4 CPUs * 64 GB	At least 4	50 to 110	7 to 16
Extra large	40,000 to 180,000 VMs Up to 2.1 million entities	4 CPUs * 96 GB	At least 5	100 to 209	10 to 20

Points to consider:

• For large deployments of Turbonomic, contact your Turbonomic representative for guidance and assistance. For extra large deployments or if you need to deploy more than one instance of Turbonomic, contact IBM Expert Labs.

• Turbonomic creates persistent volumes (PVs) based on dynamic storage provisioning. For more information, see this topic.

• The total amount of memory and CPUs required depends on the number and type of enabled mediation components and configured targets. Each target type spins up its own pod.

• Turbonomic supports namespaces with resource quotas. Contact your Turbonomic representative for help in configuring a quota.