Simplicity – Google SRE Handbook (2017)

Original Article ↗ Hacker News Discussion ↗

Trust in Google & Alleged Hypocrisy

  • Several commenters distrust Google’s advice due to:
    • Account deletions (consumer and GCP), including the high‑profile pension fund incident.
    • A perception that Google often deletes users/services once they become “inconvenient.”
  • Others argue:
    • Every large org makes mistakes; Google’s reliability is still high.
    • Guidance can be sound even if not perfectly followed internally.
    • Critiquing advice solely because Google wrote it is an ad hominem.

Applicability & Cargo Culting

  • Many see the simplicity chapter as broadly good advice for anyone running services, from solo devs to large orgs.
  • Strong warnings against:
    • Copying “Google practices” into small companies and over‑engineering.
    • Treating the SRE book as a universal manual rather than context‑specific essays.
  • Some think the content is vague, feel‑good truisms that mainly fuel Google cargo culting and LinkedIn‑style virtue signaling.

Simplicity, Complexity, and Human Factors

  • Debate over the book’s claim that emotional attachment to code drives complexity:
    • Critics: main drivers are incentives, underinvestment in maintenance, difficulty measuring long‑term costs; blaming “emotions” becomes a thought‑terminating cliché.
    • Supporters: emotional attachment, sunk cost, job security, and pride in complex systems are real and common.
  • General agreement that:
    • Complexity often accumulates via endless features and weak cleanup incentives.
    • Simplicity requires continuous pruning, not rare giant rewrites.
  • Discussion on:
    • Overuse of feature flags/configs vs deleting code.
    • Commented‑out “dead code” vs relying on version control; some find VCS‑based reversion hard in practice.

SRE Role, Culture, and Org Dynamics

  • Some portray SRE organizations as:
    • Philosophical, heavy‑handed, and sometimes contemptuous of developers.
    • Mandating complexity themselves while criticizing SWE systems.
  • Others note:
    • SRE concepts (SLOs, error budgets, etc.) are metric‑focused, but metrics are not automatically “science.”
    • Many companies label ops teams as “SRE” without giving them actual engineering authority, leading to dysfunction.
  • Observations that internal politics (SRE as “guardian of simplicity” vs SWE as “complexity drivers”) can be unhealthy.

Google Cloud Incident & Backups

  • Disagreement over the pension fund outage:
    • Some assert Google deleted the entire account and all backups, recovery only possible via another provider.
    • Others cite Google and customer statements saying GCS backups in the same region were intact and used, and that only VM configuration was deleted.
  • Downtime of ~13 days is seen by some as incompatible with “rapid” restoration claims.