px -deploy implies a 2Gi limit and request -> is limit sufficient (for clusters with large and small nodes)

[entlein]

Describe the Task
current default setup of auto-scaling GKE has very different node sizes, however PEM requires more than 1Gi Mem on the large nodes, smaller amount of Mem on the small nodes

Problem: px -deploy is almost impossible to get right since the default value is limit=2G and request=2G

• Test if removing the request causes issues -> no
• Set the request to a smaller RAM size or even remove it and the let auto-scaler figure it out . Limits must stay
• Alernatively figure out a good way to pass those values