Cron, Systemd Timers, and Automation • Software Engineering Learning Paths

Scheduled Task Execution#

Automation means running scripts without human intervention — on a schedule, in response to events, or as part of a pipeline.

flowchart TD
    Trigger["Trigger"]
    Trigger --> Cron["Cron (time-based)"]
    Trigger --> Systemd["Systemd Timer (time-based, managed)"]
    Trigger --> Event["Event (file change, webhook)"]
    Trigger --> CI["CI/CD Pipeline (code push)"]
    
    Cron --> Script["Script Execution"]
    Systemd --> Script
    Event --> Script
    CI --> Script

Cron#

The traditional Unix scheduler. Runs commands at specified times.

Crontab Format#

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

Common Schedules#

Schedule	Cron Expression	Meaning
Every minute	`* * * * *`	Testing only
Every 5 minutes	`/5 * * *`	Health checks
Hourly	`0 * * * *`	Log rotation
Daily at 2 AM	`0 2 * * *`	Backups
Weekdays at 9 AM	`0 9 * * 1-5`	Reports
Monthly (1st)	`0 0 1 * *`	Cleanup
Every Sunday	`0 0 * * 0`	Weekly maintenance

Managing Crontab#

crontab -e              # edit your crontab
crontab -l              # list your crontab
crontab -r              # remove your crontab (careful!)
sudo crontab -u alice -e  # edit another user's crontab

Cron Environment#

Critical: Cron runs with a minimal environment — NOT your interactive shell:

No .bashrc or .bash_profile loaded
Minimal PATH (usually just /usr/bin:/bin)
No terminal (no interactive input)
Working directory is the user’s home

# BAD — relies on user's PATH
*/5 * * * * health-check.sh

# GOOD — absolute paths
*/5 * * * * /opt/scripts/health-check.sh

# GOOD — set PATH in script
*/5 * * * * /opt/scripts/health-check.sh

Writing Cron-Safe Scripts#

#!/bin/bash
set -euo pipefail

# 1. Set PATH explicitly
PATH="/usr/local/bin:/usr/bin:/bin"
export PATH

# 2. Redirect output to log file
exec >> /var/log/myapp/backup.log 2>&1

# 3. Timestamp every run
echo "=== $(date '+%Y-%m-%d %H:%M:%S') ==="

# 4. Lock file — prevent overlapping runs
LOCKFILE="/tmp/backup.lock"
if ! mkdir "$LOCKFILE" 2>/dev/null; then
    echo "Already running (lock exists: $LOCKFILE)"
    exit 0
fi
trap 'rmdir "$LOCKFILE"' EXIT

# 5. Actual work
echo "Starting backup..."
/usr/bin/rsync -a /data/ /backup/data/
echo "Backup complete"

Cron Output and Notifications#

By default, cron emails output to the user. Control this:

# Discard all output
0 2 * * * /opt/scripts/backup.sh > /dev/null 2>&1

# Log to file
0 2 * * * /opt/scripts/backup.sh >> /var/log/backup.log 2>&1

# Email only on error
0 2 * * * /opt/scripts/backup.sh > /dev/null
# (stderr still goes to email)

# Set email recipient
MAILTO="ops@example.com"
0 2 * * * /opt/scripts/backup.sh

Systemd Timers (Modern Alternative)#

Systemd timers offer advantages over cron:

Feature	Cron	Systemd Timer
Logging	Manual	journalctl
Missed runs	Lost	`Persistent=true`
Dependencies	None	Full systemd deps
Resource limits	None	cgroups
Randomized delay	None	`RandomizedDelaySec`
Status monitoring	`crontab -l`	`systemctl list-timers`

Timer Unit#

# /etc/systemd/system/backup.timer
[Unit]
Description=Daily backup timer

[Timer]
OnCalendar=*-*-* 02:00:00
Persistent=true
RandomizedDelaySec=300

[Install]
WantedBy=timers.target

Service Unit#

# /etc/systemd/system/backup.service
[Unit]
Description=Backup service
After=network-online.target

[Service]
Type=oneshot
ExecStart=/opt/scripts/backup.sh
User=backup
StandardOutput=journal
StandardError=journal

Managing Timers#

# Enable and start
systemctl enable --now backup.timer

# Check status
systemctl list-timers --all
systemctl status backup.timer
systemctl status backup.service

# View logs
journalctl -u backup.service --since today

# Run manually (test)
systemctl start backup.service

OnCalendar Syntax#

OnCalendar=hourly
OnCalendar=daily
OnCalendar=weekly
OnCalendar=*-*-* 02:00:00          # daily at 2 AM
OnCalendar=Mon-Fri *-*-* 09:00:00  # weekdays at 9 AM
OnCalendar=*-*-01 00:00:00         # first of month

Automation Patterns#

Watchdog (Restart on Failure)#

#!/bin/bash
set -euo pipefail

SERVICE="myapp"
PIDFILE="/var/run/${SERVICE}.pid"

if [[ -f "$PIDFILE" ]] && kill -0 "$(cat "$PIDFILE")" 2>/dev/null; then
    exit 0  # already running
fi

echo "$(date): $SERVICE is down, restarting..." >> /var/log/watchdog.log
/opt/scripts/start-${SERVICE}.sh

Log Rotation#

#!/bin/bash
set -euo pipefail

LOG_DIR="/var/log/myapp"
MAX_AGE=30  # days

find "$LOG_DIR" -name "*.log" -mtime +$MAX_AGE -delete
find "$LOG_DIR" -name "*.log" -size +100M -exec gzip {} \;

Health Check with Alerting#

#!/bin/bash
set -euo pipefail

ENDPOINTS=(
    "http://api.example.com/health"
    "http://web.example.com/health"
)

for url in "${ENDPOINTS[@]}"; do
    if ! curl -sf --max-time 5 "$url" > /dev/null; then
        echo "ALERT: $url is DOWN" | mail -s "Health Check Failed" ops@example.com
    fi
done

Key Takeaways#

Cron has minimal environment — always set PATH, use absolute paths
Lock files prevent overlapping — use mkdir for atomic locking
Log everything — redirect output, add timestamps
Systemd timers > cron for modern systems (logging, dependencies, missed runs)
Persistent=true catches up on missed runs after downtime
Test cron scripts manually first — env -i bash script.sh simulates cron’s environment
Never assume interactive environment in automated scripts