What it is
A Pending pod with a FailedScheduling event means the kube-scheduler ran every filter on every node and none passed. The message counts nodes per reason, for example 1 Insufficient cpu, 2 node(s) had untolerated taint.
When to use it
Use it when a Pod, Deployment or StatefulSet will not start. Paste the manifest, your nodes and optionally the pods already running, and read which filter rejected each node and what each rejection means.
Gotchas & limits
The scheduler stops at the first failing filter per node, so fixing one reason can reveal another; the suggested fixes are re-run until every pod is placed. This is an educational model of Kubernetes v1.37 scheduling, not your live cluster: autoscalers, webhooks and custom schedulers are not included.