etcd version:

{"etcdserver":"3.3.10","etcdcluster":"3.3.0"}

My understanding is that etcd_server_slow_apply_total is related to add, delete or update on disk that is handled slowly.

We didn't do anything Kubernetes and etcd; the pod didn't restart. Why did delta(etcd_server_slow_apply_total{project="XX"}[3m]) increase to 40?

See this screen shot for illustration: