TIL not to use resource reservations on frequent Kubernetes CronJobs

If you have a Kubernetes CronJob that runs every minute and a job fails, the failed pod sticks around with its resource reservations still held. With a new pod spawning every minute, your cluster resources can be exhausted very quickly.

Consider either removing resource requests from frequent CronJobs or setting a strict failedJobsHistoryLimit:

apiVersion: batch/v1
kind: CronJob
metadata:
  name: my-job
spec:
  schedule: "* * * * *"
  failedJobsHistoryLimit: 1
  jobTemplate:
    spec:
      template:
        spec:
          containers:
            - name: my-job
              image: my-image
              # no resources.requests here

Without this, a single downstream outage can cascade into cluster-wide resource pressure just from piling up failed pods.