Crontab Examples for Self-Hosting

What Is Crontab?

Crontab is the configuration file for cron, Linux’s built-in task scheduler. Every self-hosted server needs automated maintenance — backups, log rotation, certificate renewal, container cleanup. Crontab handles all of it.

If you need a refresher on cron syntax, read Cron Jobs for Maintenance first. This guide focuses on practical, copy-paste-ready examples for self-hosting.

Prerequisites

A Linux server with SSH access
Basic Linux command line knowledge
cron installed (present on virtually all Linux distributions)

Crontab Syntax Refresher

┌───────── minute (0-59)
│ ┌─────── hour (0-23)
│ │ ┌───── day of month (1-31)
│ │ │ ┌─── month (1-12)
│ │ │ │ ┌─ day of week (0-7, 0 and 7 = Sunday)
│ │ │ │ │
* * * * * command

Edit your crontab with:

crontab -e

Backup Examples

Daily Docker Volume Backup

Back up all named Docker volumes to a timestamped tarball every night at 2 AM:

0 2 * * * /usr/bin/docker run --rm -v backup_data:/data -v /backups:/backup alpine tar czf /backup/volumes-$(date +\%Y\%m\%d).tar.gz -C /data . >> /var/log/backup.log 2>&1

Weekly Database Dump

Export a PostgreSQL database from a Docker container every Sunday at 3 AM:

0 3 * * 0 /usr/bin/docker exec postgres-db pg_dump -U postgres myapp > /backups/db/myapp-$(date +\%Y\%m\%d).sql 2>> /var/log/backup.log

For MySQL/MariaDB:

0 3 * * 0 /usr/bin/docker exec mariadb-server mysqldump -u root -p"${MYSQL_ROOT_PASSWORD}" --all-databases > /backups/db/all-$(date +\%Y\%m\%d).sql 2>> /var/log/backup.log

Restic Backup Every 6 Hours

If you use restic for incremental backups:

0 */6 * * * /usr/bin/restic -r /mnt/backup-drive/restic-repo backup /srv/data --exclude-caches --tag automated >> /var/log/restic.log 2>&1

Backup Rotation — Delete Files Older Than 30 Days

30 3 * * * find /backups -name "*.tar.gz" -mtime +30 -delete >> /var/log/backup-cleanup.log 2>&1

Docker Maintenance

Prune Unused Docker Resources Weekly

Dangling images, stopped containers, and unused networks waste disk space:

0 4 * * 0 /usr/bin/docker system prune -af --volumes >> /var/log/docker-prune.log 2>&1

Remove the --volumes flag if you want to keep unused volumes.

Prune Only Dangling Images Daily

Less aggressive — only removes untagged images:

0 4 * * * /usr/bin/docker image prune -f >> /var/log/docker-prune.log 2>&1

Restart a Container Every Night

Some apps develop memory leaks. Restart them during low-traffic hours:

0 5 * * * /usr/bin/docker restart jellyfin >> /var/log/container-restart.log 2>&1

Check for Docker Image Updates

Using Diun or a simple script:

0 8 * * * /usr/bin/docker pull ghcr.io/immich-app/server:v1.99.0 2>&1 | grep -q "Status: Downloaded newer image" && echo "$(date): immich update available" >> /var/log/updates.log

SSL Certificate Renewal

Certbot Auto-Renewal

Certbot installs its own cron/systemd timer, but if you need to set it up manually:

0 0,12 * * * /usr/bin/certbot renew --quiet --deploy-hook "docker restart nginx-proxy" >> /var/log/certbot.log 2>&1

Runs twice daily (Let’s Encrypt recommends this). Only renews certificates within 30 days of expiry.

Monitoring and Health Checks

Check if a Container Is Running Every 5 Minutes

*/5 * * * * /usr/bin/docker ps --filter "name=nextcloud" --filter "status=running" -q | grep -q . || (/usr/bin/docker start nextcloud && echo "$(date): Restarted nextcloud" >> /var/log/container-health.log)

Disk Space Alert

Send a notification when disk usage exceeds 85%:

0 */4 * * * df -h / | awk 'NR==2 {gsub(/%/,"",$5); if($5 > 85) print strftime("%Y-%m-%d %H:%M") " ALERT: Disk usage at " $5 "%"}' >> /var/log/disk-alert.log

Check HTTP Endpoint Health

Verify your services are responding:

*/10 * * * * curl -sf --max-time 10 https://cloud.yourdomain.com/status.php > /dev/null || echo "$(date): Nextcloud DOWN" >> /var/log/health-check.log

Log Management

Compress Logs Older Than 7 Days

0 1 * * * find /var/log -name "*.log" -mtime +7 -not -name "*.gz" -exec gzip {} \; 2>/dev/null

Truncate Large Log Files Monthly

For application logs that grow unbounded:

0 0 1 * * for f in /srv/*/logs/*.log; do [ $(stat -f%z "$f" 2>/dev/null || stat -c%s "$f") -gt 104857600 ] && truncate -s 0 "$f" && echo "$(date): Truncated $f" >> /var/log/log-rotation.log; done

This truncates any log file over 100 MB.

System Maintenance

Update Package Lists Weekly

0 6 * * 1 apt-get update >> /var/log/apt-update.log 2>&1

Don’t auto-upgrade — just update the package index so you know what’s available. Review and apply updates manually or with unattended-upgrades for security patches only.

Security Updates Only — Automatic

On Debian/Ubuntu with unattended-upgrades configured:

0 6 * * * /usr/bin/unattended-upgrade -d >> /var/log/unattended-upgrades/cron.log 2>&1

Sync Time with NTP

Most modern systems use systemd-timesyncd, but if you need manual sync:

0 */6 * * * /usr/sbin/ntpdate pool.ntp.org >> /var/log/ntp-sync.log 2>&1

Dynamic DNS Update

If your home IP changes and you use Dynamic DNS:

*/5 * * * * /usr/local/bin/update-ddns.sh >> /var/log/ddns.log 2>&1

Crontab Best Practices

Always Redirect Output

Without redirection, cron sends output as email (which usually fails on servers without a mail system configured):

# Bad — output goes nowhere useful
0 2 * * * /backup.sh

# Good — output goes to a log file
0 2 * * * /backup.sh >> /var/log/backup.log 2>&1

Use Full Paths

Cron runs with a minimal PATH. Always use absolute paths for commands:

# Bad — docker might not be in cron's PATH
0 2 * * * docker system prune -f

# Good
0 2 * * * /usr/bin/docker system prune -f

Find the full path with which docker.

Escape Percent Signs

In crontab, % is a special character (newline). Escape it with \%:

# Bad — breaks
0 2 * * * echo $(date +%Y%m%d)

# Good
0 2 * * * echo $(date +\%Y\%m\%d)

Stagger Jobs

Don’t schedule everything at the same time:

# Bad — everything at 2 AM
0 2 * * * /backup.sh
0 2 * * * /docker-prune.sh
0 2 * * * /cert-renew.sh

# Good — spread across the night
0 2 * * * /backup.sh
30 3 * * * /docker-prune.sh
0 4 * * * /cert-renew.sh

Test Before Scheduling

Run the command manually first to verify it works:

# Test the command
/usr/bin/docker system prune -af --volumes

# Then add to crontab
crontab -e

Common Mistakes

Job Runs But Nothing Happens

Most common cause: wrong PATH. Cron’s default PATH is minimal. Use full paths to all binaries.

Job Doesn’t Run At All

Check cron is running:

systemctl status cron

Check the cron log:

grep CRON /var/log/syslog

Verify your crontab is saved:

crontab -l

“No MTA installed” Warnings

Cron tries to email command output. Either redirect output to a file (recommended) or install a mail system.

Timezone Issues

Cron uses the system timezone. Check with timedatectl. If your server is in UTC but you want jobs at 2 AM local time, convert the hour.

Next Steps

Crontab covers most scheduling needs for self-hosting. For more complex scheduling (dependencies between jobs, retry logic, conditional execution), consider systemd timers or self-hosted automation tools like n8n.

FAQ

Should I use cron or systemd timers?

Cron for simple, recurring jobs. Systemd timers for jobs that need dependency management, better logging, or resource controls. For most self-hosting maintenance tasks, cron is simpler and sufficient.

How do I see what cron jobs are scheduled?

Run crontab -l for your user’s jobs. Check /etc/crontab and /etc/cron.d/ for system-level jobs. Some packages install their own cron jobs in /etc/cron.daily/, /etc/cron.weekly/, etc.

Can I run a cron job as a specific user?

Yes. Edit that user’s crontab with sudo crontab -u username -e. Or use the system crontab /etc/crontab which has a user field.

How do I debug a failing cron job?

Redirect both stdout and stderr to a log file (>> /var/log/myjob.log 2>&1), then check the log after the scheduled time. Also check /var/log/syslog for cron execution records.