Mentoring Juniors Through Their First On-Call Rotation
A junior's first month on-call is the highest-leverage learning of their career. Done well, they end the month confident. Done poorly, they end it terrified and stay junior longer than they should.
The weight of first on-call
The first time a junior's pager goes off at 2am, the feedback loop they get from that incident shapes their relationship with on-call for years. Plan it.
The investment's leverage. A well-mentored first month produces an SRE who's confident, asks good questions, and stays in the field. A poorly-mentored first month produces an SRE who fears the pager, hides mistakes, and leaves within 18 months. The difference shows up in retention metrics 2-3 years out.
The cultural signal. How seniors handle juniors' first on-call shapes the team's broader on-call culture. Patient, instructive seniors signal that on-call is learnable; impatient, blame-prone seniors signal that on-call is a hazing ritual. Most teams underestimate how visible this signal is.
Phase 1: shadow (2 weeks)
Junior is on the rotation but their primary is the senior. Junior is paged second; they observe; they ask questions in the channel; the senior explains. Goal: the junior sees the rhythm of an on-call shift before they own one.
The mechanics. The junior is added to the rotation as "shadow primary"; the actual primary is a senior. Pages go to the senior; the junior sees them, observes the response, asks questions in a separate channel. The senior narrates what they're doing and why.
What the junior learns in shadow. The pace of investigation. How seniors form hypotheses. When to escalate vs. continue solo. The conversations during incident response that aren't in any runbook. The cultural rhythm of on-call.
Phase 2: pair (2 weeks)
Junior is primary; senior is secondary and on the same call. Junior drives; senior comments only when the junior is stuck or going visibly wrong. Goal: the junior feels the pressure of being primary while still having a safety net.
The discipline of "comments only when stuck." Senior wants to take over; resist. The junior's growth requires them to drive even when slower than the senior would be. The senior's job is to coach, not perform.
The signals to step in. Junior is heading toward irreversible action without considering alternatives ("I'm going to delete the index"). Junior has been stuck for 10+ minutes with no progress. Junior has visibly missed a critical detail. Each justifies senior intervention; ambiguous cases default to letting the junior continue.
Phase 3: solo (2 weeks supervised)
Junior is alone. Senior reviews the timeline of every incident and gives feedback the next day. Goal: the junior owns the call; learning continues with a delay.
The asynchronous review. The senior isn't on the bridge; they review the postmortem and the channel transcript afterwards. The feedback is detailed: what went well, what to do differently, what specific moves to practice next time. The junior can act on it during their next shift.
The discipline of not interrupting. Senior sees a question in the channel; resists the urge to answer. The junior's responsibility includes asking the right people; if the senior intervenes, the junior doesn't learn the escalation muscle.
What to coach
Hypothesis formation. Communication clarity ("what is the customer impact right now?"). When to escalate. When to stop debugging and just roll back. None of these are technical skills; all of them are what makes the difference between a junior who grows fast and one who grinds slowly.
The coaching specifics. Hypothesis formation: ask "what are 3 different things this could be?" before letting the junior commit to one. Communication: have them write the customer comm before the team has the fix; the writing exercise reveals confused understanding. Escalation: practice escalation on low-stakes incidents so the muscle is there for high-stakes.
The "rollback over fix" coaching. Juniors over-commit to fixing forward because they want to demonstrate technical skill. Senior coaches: "rollback is a valid first response; we can fix forward tomorrow." The lesson is that incident response is about reducing customer impact, not about engineering pride.
Two failure modes to avoid
Over-coddling: the senior takes the call from the junior at the first sign of stress, and the junior never owns an incident. Under-coverage: the junior is technically alone before they are ready, and they handle a bad incident badly, and they fear the pager for a year. The phased schedule above avoids both.
The over-coddling pathology. Junior gets paged; struggles for 10 minutes; senior takes over. Junior never experiences the satisfaction of resolving an incident on their own. Confidence doesn't build. After 6 months, they're still not running incidents solo.
The under-coverage pathology. Junior is on rotation alone too soon. They handle a SEV2 badly; the postmortem is rough; their 1:1 with manager goes poorly. They associate on-call with embarrassment. They quit or transfer. The shadow → pair → solo phasing is what avoids this.
Common antipatterns
The "throw them in the deep end" mentor. Senior thinks struggle is character-building. Sometimes is; usually produces fearful juniors. Phased mentoring works better.
The "I'll fix it then explain" approach. Senior takes over to resolve faster; explains afterwards. Junior doesn't learn during the actual incident. Coach during; explain after.
No formal mentor for the rotation. Junior joins on-call; nobody is responsible for their growth. The team assumes "they'll figure it out." Most don't, fast. Always assign a senior mentor for the first 8-12 weeks.
The mentor who doesn't post-review. Solo phase but no review of the junior's incidents. The junior gets no feedback; quality drifts. The 30-minute next-day review is what makes solo phase a teaching moment.
What to do this week
Three moves. (1) For your team's next new on-caller, formalise the shadow → pair → solo schedule. Most teams have an informal version; explicit phases reduce ambiguity. (2) Pair each junior with a specific senior mentor for the 6-week period. The accountability is what makes the mentoring real. (3) Schedule the post-shift reviews on the senior's calendar for the 2-week solo phase. Without scheduled reviews, they don't happen.