K8s CronJob: Advanced Scheduling Patterns

💡 Quick Answer: CronJob uses standard cron syntax: schedule: "0 2 * * *" (2 AM daily). Key settings: concurrencyPolicy: Forbid (skip if previous still running), startingDeadlineSeconds: 300 (skip if >5min late), successfulJobsHistoryLimit: 3, failedJobsHistoryLimit: 3. Timezone support (K8s 1.27+): timeZone: "America/New_York". Suspend without deleting: spec.suspend: true.

The Problem

Scheduled tasks in Kubernetes need:

• Reliable cron-like scheduling
• Handling of overlapping executions
• Failure detection and retry policies
• History management (don’t fill etcd with old Jobs)
• Timezone-aware scheduling

The Solution

Full CronJob Configuration

apiVersion: batch/v1
kind: CronJob
metadata:
  name: daily-backup
  namespace: production
spec:
  schedule: "0 2 * * *"              # 2 AM daily
  timeZone: "Europe/Rome"            # K8s 1.27+ timezone support
  concurrencyPolicy: Forbid          # Skip if previous still running
  startingDeadlineSeconds: 300       # Skip if >5 min late
  suspend: false                     # Set true to pause
  successfulJobsHistoryLimit: 3      # Keep last 3 successful Jobs
  failedJobsHistoryLimit: 5          # Keep last 5 failed Jobs
  
  jobTemplate:
    spec:
      backoffLimit: 3                # Retry failed jobs 3 times
      activeDeadlineSeconds: 3600    # Kill if running >1 hour
      ttlSecondsAfterFinished: 86400 # Auto-delete after 24h
      template:
        spec:
          restartPolicy: OnFailure
          serviceAccountName: backup-sa
          containers:
          - name: backup
            image: backup-tool:v2
            command: ["/backup.sh"]
            env:
            - name: S3_BUCKET
              valueFrom:
                secretKeyRef:
                  name: backup-secrets
                  key: bucket
            resources:
              requests:
                cpu: 500m
                memory: 512Mi
              limits:
                cpu: "2"
                memory: 2Gi

Cron Schedule Syntax

┌───────────── minute (0-59)
│ ┌───────────── hour (0-23)
│ │ ┌───────────── day of month (1-31)
│ │ │ ┌───────────── month (1-12)
│ │ │ │ ┌───────────── day of week (0-6, Sun=0)
│ │ │ │ │
* * * * *

Examples:
"*/5 * * * *"      Every 5 minutes
"0 * * * *"        Every hour
"0 2 * * *"        Daily at 2 AM
"0 9 * * 1"        Monday at 9 AM
"0 0 1 * *"        First of every month
"0 */6 * * *"      Every 6 hours
"30 8 * * 1-5"     Weekdays at 8:30 AM
"0 0 * * 0"        Weekly on Sunday midnight

Concurrency Policies

# Allow — default, multiple Jobs can run simultaneously
concurrencyPolicy: Allow
# Previous Job still running + new schedule → BOTH run

# Forbid — skip new if previous still running
concurrencyPolicy: Forbid
# Previous Job still running + new schedule → new SKIPPED

# Replace — kill previous, start new
concurrencyPolicy: Replace
# Previous Job still running + new schedule → previous KILLED, new starts

Manage CronJobs

# List CronJobs
kubectl get cronjob -n production
# NAME           SCHEDULE      SUSPEND   ACTIVE   LAST SCHEDULE
# daily-backup   0 2 * * *     False     0        2h ago

# Trigger manually (create Job from CronJob)
kubectl create job --from=cronjob/daily-backup manual-backup-001

# Suspend (pause without deleting)
kubectl patch cronjob daily-backup -p '{"spec":{"suspend":true}}'

# Resume
kubectl patch cronjob daily-backup -p '{"spec":{"suspend":false}}'

# View Job history
kubectl get jobs -l job-name -n production --sort-by=.status.startTime

# Check last Job logs
kubectl logs job/daily-backup-28456320 -n production

# Delete CronJob (also deletes owned Jobs)
kubectl delete cronjob daily-backup

Common Patterns

# Pattern 1: Database cleanup (skip if previous still running)
apiVersion: batch/v1
kind: CronJob
metadata:
  name: db-cleanup
spec:
  schedule: "0 3 * * *"
  concurrencyPolicy: Forbid
  jobTemplate:
    spec:
      template:
        spec:
          restartPolicy: Never
          containers:
          - name: cleanup
            image: postgres:16
            command:
            - psql
            - -h
            - $(DB_HOST)
            - -U
            - $(DB_USER)
            - -c
            - "DELETE FROM logs WHERE created_at < NOW() - INTERVAL '30 days'"

---
# Pattern 2: Report generation with email
apiVersion: batch/v1
kind: CronJob
metadata:
  name: weekly-report
spec:
  schedule: "0 9 * * 1"           # Monday 9 AM
  timeZone: "Europe/Rome"
  jobTemplate:
    spec:
      backoffLimit: 2
      template:
        spec:
          restartPolicy: OnFailure
          containers:
          - name: report
            image: report-gen:v3
            env:
            - name: REPORT_TYPE
              value: weekly
            - name: EMAIL_TO
              value: team@example.com

---
# Pattern 3: Certificate renewal check
apiVersion: batch/v1
kind: CronJob
metadata:
  name: cert-check
spec:
  schedule: "0 8 * * *"           # Daily 8 AM
  successfulJobsHistoryLimit: 1
  failedJobsHistoryLimit: 3
  jobTemplate:
    spec:
      template:
        spec:
          restartPolicy: Never
          containers:
          - name: check
            image: bitnami/kubectl:1.30
            command:
            - sh
            - -c
            - |
              for ns in $(kubectl get ns -o name); do
                kubectl get secret -n ${ns##*/} -o json | \
                  jq -r '.items[] | select(.type=="kubernetes.io/tls") | .metadata.name'
              done

Common Issues

CronJob not triggering

Check startingDeadlineSeconds — if controller was down and missed the window, the Job is skipped. Also check suspend: false.

Too many Jobs accumulating

Set successfulJobsHistoryLimit and failedJobsHistoryLimit. Also use ttlSecondsAfterFinished on the Job spec.

Job runs twice at the same time

Default concurrencyPolicy: Allow. Change to Forbid or Replace.

Timezone not working

Requires K8s 1.27+. Check: kubectl version. Older versions use UTC only.

Best Practices

• concurrencyPolicy: Forbid for most jobs — prevent overlap
• Set startingDeadlineSeconds — avoid running stale schedules after downtime
• Keep history limits low — 3-5 prevents etcd bloat
• restartPolicy: Never for debugging (keep failed pods for log inspection)
• restartPolicy: OnFailure for automatic retries within a Job
• Monitor CronJob execution — alert on missed schedules or failures

Key Takeaways

• CronJobs create Jobs on a cron schedule — standard 5-field syntax
• Concurrency policies: Allow (default), Forbid (skip), Replace (kill+restart)
• startingDeadlineSeconds prevents stale executions after missed schedules
• Set history limits and TTL to prevent resource accumulation
• Timezone support in K8s 1.27+ — no more UTC-only scheduling