Cookie Preferences
We use cookies to enhance your experience, analyse site traffic, and support our marketing. You can customise your preferences or accept all cookies. See our Privacy & Cookie Policy for details.

DevOps & Platform Engineering

The On-Call Rotation That Doesn't Burn Out Your Best Engineer

DevOps & Platform Engineering

6 min read

The most competent engineer on a team is often the one who gets paged the most, informally, regardless of the official rotation — because colleagues escalate to whoever they trust to fix it fastest. That pattern quietly burns out exactly the person you can least afford to lose.

A sustainable on-call design has to account for this informal escalation pattern explicitly, not just publish a fair-looking rota and assume the problem is solved.

Why a technically fair rota still produces unfair burnout

An evenly rotated schedule looks fair on paper while informal escalation patterns route the hardest pages to the same one or two people every time, regardless of whose week it officially is. The rota's fairness is cosmetic if the actual incident load isn't evenly distributed underneath it.

Measuring who actually resolves incidents, not just who's officially on call, reveals this gap quickly once anyone looks for it.

What a genuinely sustainable design includes

Runbooks detailed enough that any on-call engineer, not just the most senior one, can resolve common incidents independently, an explicit policy discouraging off-rota escalation to specific individuals, and real compensating time off after a disruptive on-call week, are what actually protect the person currently absorbing the informal load.

A fair-looking rota with an unfair informal escalation pattern underneath it isn't fair. It's invisible until the best engineer quits.MNB DevOps Practice

Key actions

  • Measure who actually resolves incidents, not just who's officially on call
  • Build runbooks detailed enough for any on-call engineer to use independently
  • Discourage off-rota escalation to specific individuals explicitly
  • Provide real compensating time off after disruptive on-call weeks

Discuss this with an architect →