SLO Handoff Between Teams

Service ownership moves; SLOs that do not move with it become orphans. The handoff is mechanical but rarely done.

Why SLOs orphan

Team A owns service X with an SLO; team B inherits service X; the SLO stays in A’s alerts and A’s on-call gets paged for B’s service. The SLO must move; in practice it forgets to.

• Service moves, SLO doesn’t. Service catalog updates; alert routing keeps pointing at the previous owner.
• Wrong team paged. A’s on-call wakes up for B’s outage; cannot fix it; the page is noise plus delay.
• No catalog audit. Without a quarterly check, mismatches accumulate; the alerts become advisory then ignored.
• The fix. A four-step handoff: document, route, dashboard ownership, joint period. Mechanical, but rarely done.

Four-step handoff

• 1. Document the SLO and its history.
• 2. Update alert routing to new team.
• 3. Update dashboard ownership.
• 4. Joint owners during transition.

Joint-ownership transition

Old team and new team jointly own the service for one quarter. Old team mentors; new team learns the SLO’s context. Cleaner than instant handoff; preserves institutional memory.

• Old team mentors. Walk the new team through historical incidents, SLO threshold reasoning, known-flaky alerts.
• New team shadows. Joins the old team’s on-call rotation for one quarter; sees the SLO in motion.
• Joint pages. Both teams paged on alert during the transition; old team can disengage when comfortable.
• Cutover. End of quarter, alert routing flips fully to new team; old team archived as escalation contact.

Quarterly audit

Quarterly review: list every service, its owning team, and the team SLO alerts route to. Mismatches get re-handed off explicitly. Without the audit, drift accumulates.

• Service catalog. Source of truth for ownership; every service has a current owner; no orphans.
• Alert routing. PagerDuty / Opsgenie configuration verified against the catalog; mismatches flagged.
• Dashboard ownership. Who fixes a broken Grafana board? Same team as the alert; mismatch is a smell.
• Quarterly cadence. Slow enough that the audit is not busywork; fast enough that drift stays bounded.

Antipatterns

• Silent handoff. Old team continues to own (or nobody does).
• Instant handoff with no joint period. New team flying blind.
• No catalog audit. Mismatches accumulate.

What to do this week

Three moves. (1) Apply the pattern to your most-impactful service. (2) Measure adherence for 30 days. (3) Rewrite the policy or the SLO if the gap is durable.