Supporting material
Desk-researched interview 4: Can a newcomer operate a service?
Desk research, not original interviews. These question sets synthesize published accounts; no practitioner was interviewed for this book.
Question: Can a new engineer find a service owner, run its tests, and explain a safe rollback without asking a particular veteran?
Method: Desk research from published practitioner accounts and operating guidance; no new hire was observed attempting all three tasks. Research snapshot: November 2026.
What the published accounts say
- Rippling infrastructure engineer JD Palomino says that a previously manual hunt for the owner of failing code—including asking an “old guard”—led to a service catalog mapping file paths to teams and on-call channels. After adoption, infrastructure engineers could identify the on-call rotation within 30 seconds, and alerts were largely routed automatically. This supports owner discovery for failures by infrastructure staff, not a measured first-day success rate for new hires. It took a four-week organization-wide migration to get ownership coverage. Rippling, Code ownership at Rippling, 2024.
- Getir's former technical product manager describes a Backstage service catalog, documentation and templates. In this personal account, the reported time for new hires to reach their tenth PR went from 60 to 20 days. That is a self-reported onboarding proxy, not proof that they could independently run a service's tests, deploy it or reverse a bad release. Guifré Ballester, Implemented IDP for Faster Onboarding.
Synthesis, not testimony
Answer: unknown for the complete three-part test. Owner routing is documented; independent tests and safe rollback are not established by these sources. An ownership map is not an operating manual. A catalog can say who is paged without teaching how to reproduce a test failure or when to halt or roll back a deployment. The primary outcomes are efficient onboarding and reliable recovery; independence and learning also contribute to engaging.
Run a consenting, low-risk onboarding exercise: give a new engineer a service name and normal access, not a preferred human contact. Observe whether they identify the current owner and escalation path, get the repo running, execute a relevant test suite, find the deployment runbook and explain rollback preconditions, data migration constraints, feature flags and verification after reversal. Use a sandbox or tabletop scenario, never an unapproved production rollback. Record time, missing permissions, stale links and points requiring veteran knowledge; fix the path and repeat with another newcomer.
Questions to ask a practitioner
- Can you show the catalog entry and prove its owner and on-call links are current?
- Can a newcomer run the tests from written instructions with ordinary access? What failed on the first attempt?
- What does rollback not undo (for example, a schema migration or external side effect), and who authorizes it?
Follow-up desk-researched responses
1. Catalog entry and current owner? Rippling gives a sample billing service entry with file paths, GitHub team and on-call handles, and describes deriving alert routing from it. Backstage's documentation says owning teams maintain catalog metadata in source control. These show how an owner can be found; neither proves that a particular entry is current today or that a new hire found it unaided. Verify the displayed owner against the live on-call schedule and an actual failed-build escalation. Rippling; Backstage catalog documentation.
2. Can a newcomer run tests from written instructions? Not demonstrated by Rippling's owner-routing account or Getir's time-to-tenth-PR proxy. Google's SRE workbook provides a related training example: new team members used documentation, guided labs and starter projects before entering on-call. That is evidence for practiced onboarding with mentoring, not for an unaided run of a particular service's test suite; test commands, access failures and first-attempt success remain local observations to collect. Google SRE, On-Call.
3. Rollback limits and authority? Google SRE advises “detect, roll back, fix, and roll forward” for appropriate changes, but cautions that a rollback might not repair corrupted data and that incompatible changes can be difficult to reverse. Its onboarding example includes rollback in a hands-on training checklist and defined escalation paths. That does not establish who authorizes reversal for a given service; the service's runbook and on-call policy must answer it. Google SRE, On-Call.
Evidence gap: No cited source observes one new engineer completing all three tasks unaided. “Time to tenth PR” and “owner found in 30 seconds” are different measures.