contour_memory_request
Memory request for the Contour control-plane pods.
Description
The contour_memory_request parameter sets the Kubernetes memory request applied to the Contour control-plane pods on a Rack. It takes effect when router_type=contour. The Contour control plane watches your routing configuration and translates it into Envoy configuration, so the amount of memory it needs grows with the number of routes the Rack serves.
On a Rack with many routes, the control plane holds a larger configuration in memory. Raising contour_memory_request reserves enough memory for that configuration and keeps the control plane from running low under load.
This parameter has no effect on Racks running router_type=nginx. It applies only to AWS Racks using Contour.
Default Value
The default value is 128Mi. This default preserves existing behavior, so Racks that do not set the parameter run the Contour control plane with the standard memory request.
Use Cases
- Raise the value on Racks with many routes, where the Contour control plane holds a larger configuration and can otherwise run low on memory.
- Keep the default on Racks with a small number of routes, where
128Miis sufficient for the control plane.
Setting the Parameter
$ convox rack params set contour_memory_request=256Mi -r rackName
Setting parameters... OK
Viewing Current Configuration
$ convox rack params -r rackName
Additional Information
This parameter is clearable. Clearing it returns the control plane to the default request of 128Mi:
$ convox rack params set contour_memory_request= -r rackName
Setting parameters... OK
The value is a Kubernetes memory quantity (for example 128Mi, 256Mi, or 512Mi). It sets a request, not a limit, so it reserves scheduling capacity rather than capping usage. The parameter applies only when router_type=contour; on router_type=nginx Racks it has no effect regardless of its value. It is available on AWS Racks only.
See Also
