![[Pasted image 20250206064444.png|scheduler-arch]] `scheduler` handles the **pod-to-node** process. This process is done by **filtering** and **scoring**. 1. Filtering is when it looks for the requirements of pod in each nodes (e.g. a pod needs at least 4 CPU cores). 2. In scoring, it scores the node based of the specification (e.g. data locality). 3. A random node(s) of high scores is selected. ![[Pasted image 20250206065236.png]]