Federating reliability ownership to product teams

WHAT THIS TESTS: Whether you recognize that a central SRE team that does all reliability work cannot scale, and that the fix is to shift to a platform and enablement model that distributes ownership while preserving standards.

A GOOD ANSWER COVERS: The core move is from SRE-as-doer to SRE-as-enabler. Build a self-service reliability platform with paved roads and golden paths: standardized, well-instrumented deploy pipelines, observability baked in by default, SLO definition and burn-rate alerting tooling, runbook and dashboard templates, and safe-by-default deployment (canary, rollback, feature flags). When the easy path is also the reliable path, product teams get reliability without bespoke SRE involvement. Pair the platform with training and culture: teach teams to define and own their own SLOs, run their own on-call, and conduct blameless postmortems, with SRE consulting, reviewing, and embedding only for the highest-stakes services. Establish governance that scales without gatekeeping: shared reliability standards, lightweight production readiness criteria teams can self-certify against, and error budget policies that automatically govern the velocity-reliability trade-off so SRE does not have to approve every change. SRE shifts to building the platform, setting standards, and handling the hardest problems, while accountability for each service's reliability lives with the team that owns it.

COMMON WRONG ANSWERS: Keeping SRE as a mandatory approval gate for every deploy, which just relocates the bottleneck. Throwing responsibility over the wall to teams with no platform or training to support them. Removing all central standards, leading to inconsistent, unreliable practices.

LIKELY FOLLOW-UPS: How do you avoid each team reinventing tooling? How do you keep standards consistent without gatekeeping? How do you decide which services still get embedded SRE?

ONE CONCRETE EXAMPLE: SRE builds a paved-road platform where spinning up a new service automatically provisions dashboards, alerts, an SLO template, and a canary pipeline. Product teams own their SLOs and on-call; an error budget policy auto-governs their shipping velocity. SRE reserves hands-on involvement for the few revenue-critical services, ending the bottleneck while reliability standards hold across the org.