Scheduling

WARNING

• Be very careful with these settings, as they can break your site, Memcached or MySQL cluster. Before setting up a certain rule, make sure that you have the necessary nodes available.
• All the scheduling modifications (setting up scheduling rules, tolerations, pod deletions and k8s node upgrades) might be with downtime.

Scheduling rules

The scheduling rules describe how your site should be scheduled inside the cluster, for example you can schedule your site to always run on preemptible nodes, or to preferably use nodes from a certain region etc.

For example, you can schedule your nodes to use preemptible machines in order to reduce costs. To do this, add a preemptible node , then select the option preemptible from the Node Label Selector list.

You can choose if you want your Scheduling mode to be:

• Never - Only standard nodes are chosen for scheduling WordPress pods
• Prefer - If there are preemptible nodes available, they are chosen for scheduling WordPress pods. This method is cost efficient but not highly available.
• Always - Only preemptible nodes are chosen for scheduling WordPress pods. This method is the most cost efficient but not highly available.

On Google’s documentation you can read more about what preemptible VMs and preemptible instances are.

By default, you have only the default node pool, but you can add new nodes and setup more scheduling rules.

Tolerations

Tolerations are used to ensure that the pods are not scheduled onto inappropriate nodes.

You can read more about tolerations on the Kubernetes documentation and you can also checkout these example use cases .

Scheduling Status

You can also see the Scheduling Status, which contains the list nodes where the site is currently scheduled on.