Hardware requirements
Review the hardware and resource requirements for running production Splunk Enterprise instances in pods through the Splunk Operator.
The hardware guidelines for running production Splunk Enterprise instances in pods through the Splunk Operator are the same as running Splunk Enterprise on any supported operating system and file system.
For detailed hardware requirements, see the Splunk Enterprise Reference Hardware documentation. Also follow the same guidance for Splunk Enterprise on disabling Transparent Huge Pages (THP) for the nodes in your Kubernetes cluster.
Use Kubernetes Quality of Service classes
In addition to the hardware guidelines, you can use Kubernetes Quality of Service Classes to configure CPU and memory resource allocations that map to your service level objectives.
| QoS | Summary | Description |
|---|---|---|
| Guaranteed | CPU/Mem requests = CPU/Mem limits |
When the CPU and memory requests and limits values are equal, the pod receives a QoS class of Guaranteed. Use this level of service for Splunk Enterprise production environments. |
| Burstable | CPU/Mem requests < CPU/Mem limits |
When the CPU and memory requests value is lower than the limits, the pod receives a QoS class of Burstable. This level of service is useful in a user acceptance testing (UAT) environment, where the pods run with minimum resources and Kubernetes allocates additional resources depending on usage. |
| BestEffort | No CPU/Mem requests or limits are set |
When the requests or limits values are not set, the pod receives a QoS class of BestEffort. This level of service is sufficient for testing, or a small development task. |
For examples on how to implement these QoS classes, see the Custom Resources reference.