Custom Cost
The Custom Cost section of the Settings Page allows admins to edit, review, delete and add custom cost models to their Kubernetes environments.
This functionality is especially useful for organisations running Kubernetes environments on privately hosted infrastructure, on premises or private cloud.
The Custom Cost feature supports both regular Node Types as well as GPUs.
Admins can add Node Types with custom CPU cores and RAM capacity reflecting their private infrastructure. In addition, admins can also add a Total Cost/h, CPU Cost/h and RAM Cost/h for their custom nodes. Custom GPU types with specific Models, Regions and Cost/h can also be added.
The ability to add custom node and GPU types enables Replex to provide accurate cost allocation and chargeback for privately hosted Kubernetes environments.
Add Custom Node Type
To add a new custom node, click on + Add Node Type in the top right of the Node Types section.
On the next screen, name the new node type and optionally input the Operating System and Region of the node.
Next toggle between Total Cost and Split Cost to choose whether to input CPU and RAM costs for that node combined or individually.
When entering combined costs (Total Cost) you are required to enter Total Cost/h, the Number of CPU Cores and Gigabyte of RAM for the custom node.
Alternatively, when entering cost individually (Split Cost) you are required to enter CPU Cost/h and RAM Cost/h for that node type. You can also optionally enter the Number of CPU Cores and Gigabyte of RAM.
Click Save to create the new node type.
This will create the new custom node and populate it in the Node Types section of the Custom Cost screen.
Add Multiple Node Types
Multiple node types can be added using the csv file upload functionality.
When adding multiple node types, Kubernetes admins can opt to split CPU and RAM costs for those nodes and upload the cost model using the Split Cost csv format.
Alternatively they can combine CPU and RAM costs for node types and upload the cost model using the Total Cost csv format.
Split Cost csv format for adding multiple node types:
type; | cpu_cost; | ram_cost; | cpu_cores; | ram_gb | operating_system; | region |
instance_type_a; | 0.09; | 0.37; | 4; | 16 | linux; | west-1 |
instance_type_b; | 0.117; | 0.8236; | 8; | 32 | linux; | west-2 |
Total Cost csv model for adding multiple node types:
type; | total_cost; | cpu_cores; | ram_gb | operating_system; | region |
instance_type_a; | 0.09; | 4; | 16 | linux; | west-1 |
instance_type_b; | 0.117; | 8; | 32 | linux; | west-2 |
Once the csv has been formatted correctly click Upload CSV in the top right of the Node Types section.
Click Choose File in the pop-up screen, choose the csv and click Upload.
Once uploaded, the custom node types will be populated in the Node Types section of the Custom Cost screen..
Edit or Remove Node Types
Admins can edit previously added Node Types by clicking the edit icon in front of each node.
Previously added Node Types can be deleted by clicking on the delete icon in front of each node.
Add Custom GPU Type
To add a new GPU Type, click on + Add GPU Type in the top right of the GPU Types section of the Custom Cost screen.
On the next screen, enter the Model and Cost/h of the GPU type. Optionally enter the Region of the GPU.
Click Save to create the new GPU type.
This will create the new GPU type and populate it in the GPU Types section of the Custom Cost screen.
Add Multiple GPU Types
Multiple GPU types can be added using the csv file upload functionality.
To add a new csv file, click on Upload CSV in the top right of the GPU Types section.
Click Choose File in the pop-up screen, choose the csv and click Upload.
The uploaded csv should follow the following format:
model; | cost_hourly; | region; |
nvidia-rtx-2080;0.09; | 5.5; | west-1 |
nvidia-tesla-t4; | 2.1; | west-2 |
Once uploaded the new GPU types will be populated in the GPU Types section.
Edit or Remove GPU Types
Admins can edit previously added GPU Types by clicking the edit icon in front of each GPU.
Previously added GPU Types can be deleted by clicking on the delete icon in front of each GPU.
Last updated
