Shapes

Predefined instance types and sizes

Shapes are used to define sizes of instances and to allow control over node selection for new instances. If there are high performance nodes and low performance nodes these can be placed into two instance shapes such as m2-standard and m3-standard . This will allow the user to select the shape suitable for the instance and a node will be selected based on which node has the most available CPU and memory. Shapes are associated with nodes by matching roles. This can be done for example by adding shape-m2 to the shape Roles then adding that same role to all the nodes that fit that shape.

Flexible Shapes

Flexible shapes define default processor and memory size but allow these to be overridden when creating the instance or in the pod spec.

Last updated