DEV Community

Hive80-lab
Hive80-lab

Posted on

On-Call Scheduling That Actually Works: A Practical Guide for Small Teams

#devops #sre #automation #beginners

On-call rotations are the backbone of any reliable service. But most teams I have worked with treat scheduling as an afterthought — a spreadsheet someone updates manually, forgotten until 2am when the wrong person gets paged.

The Problem with Manual On-Call

  • Coverage gaps: Nobody realizes the schedule is wrong until an incident hits
  • Burnout: The same senior engineer gets paged every night because "they know the system"
  • No escalation path: When primary on-call does not respond, nobody knows who is next
  • Holiday blindness: Nobody plans for reduced coverage during holidays

What a Good On-Call System Looks Like

1. Clear Rotation Rules

  • Primary and secondary on-call, rotating weekly
  • No engineer on-call for more than 2 consecutive weeks
  • Handoff at a consistent time (we use Monday 10am local)

2. Escalation Policy

  • Primary has 5 minutes to acknowledge
  • Secondary gets paged at 10 minutes
  • Manager notified at 15 minutes
  • All tracked automatically

3. Follow-the-Sun for Global Teams

If you have engineers across time zones, use it:

  • APAC handles incidents 00:00-08:00 UTC
  • EMEA handles 08:00-16:00 UTC
  • Americas handles 16:00-00:00 UTC
  • Each region has primary + secondary

4. Compensation and Recovery

  • On-call bonus per week
  • Day off after a night-time incident
  • No meetings the day after a major incident

Tools That Work

  • PagerDuty: Best overall, but expensive for small teams
  • Opsgenie: Good alternative, cheaper
  • Simple cron + Slack: For teams under 5 people, this works fine
  • Custom Python script: If you want full control

The Template I Use 

# On-Call Schedule — Week of [DATE]

## Primary: [Name] — [Phone] — [Slack]
## Secondary: [Name] — [Phone] — [Slack]

## Escalation:
1. Primary (5 min)
2. Secondary (10 min)
3. Engineering Manager (15 min)
4. CTO (30 min)

## Handoff:
- Monday 10:00 local
- Review pending incidents
- Update runbook with learnings
Enter fullscreen mode Exit fullscreen mode

Common Mistakes to Avoid

  1. Do not put the newest engineer as primary — they will panic and you will escalate anyway
  2. Do not skip the handoff — context is everything in incident response
  3. Do not forget about holidays — plan coverage 2 weeks in advance
  4. Do not make on-call optional — it is a responsibility, not a favor

Want ready-to-use on-call templates, escalation policies, and automation scripts? Check out our Ops Starter Kit:

🔗 Hive80 Lab — Gumroad Store

Your team deserves better than a spreadsheet at 2am. 🌙

Top comments (0)