SRE Handbook

Stefan Coetzeecreated 2026-04-25updated 2026-06-083 min read
Vault note, not reviewed against the source. Written in the knowledge vault on 2026-04-25 by models working with Stefan Coetzee and published as it stands, with private addresses, e-mail addresses and an employer name redacted. Check claims against the primary source before relying on them.

Practical SRE training material. Opinionated, experience-driven, 30+ years of operational reality.

The Manifesto

The substrate principle: SRE as Truth Verified Working -- the discipline is about claims being true, verified, working. Parent of the ten pillars; same principle applied to LLM-substrate (Receipts Standard) and people-substrate (Three-Layer Position Model, real-emotional-maturity).

SRE Manifesto -- SRE is a role, not a team. The concentric model. Standby is the differentiator.

AI-era extensions:

GitHub-render mirror: SRE/README.md

Published expressions

  • (S)uper (R)elatable (E)xperience -- public front-door definition of SRE vs DevOps vs platform; the "Grossly oversimplified" framing of what this whole framework is.

The Pillars

#PillarFocus
1ReliabilitySLIs, SLOs, SLAs, Error Budgets
2ScalabilityHorizontal/Vertical scaling, Capacity planning
3ObservabilityMonitoring, Logging, Tracing, Alerting
4Incident ManagementOn-call, Response, Postmortems
5Infrastructure as CodeTerraform, GitOps, Immutable infrastructure
6CI/CD & DeploymentProgressive delivery, Canary, Blue-green
7PerformanceLatency, Throughput, Optimization
8SecurityShift-left, Zero-trust, Secrets management
9Cost OptimizationFinOps, Right-sizing, Waste elimination
10Toil ReductionAutomation, Self-service, Elimination

Supporting Content

  • Patterns -- reusable architecture patterns. 7 atoms covering memory architecture, retrieval, hardware/inference trade-offs.
  • Runbooks -- operational procedures. 5 canonical runbooks (outage response, DB failover, rollback deployment, cert rotation, credential rotation). 6 more planned.
  • Tools -- tool-specific guides. 3 atoms (Headscale, OpenCode, Dolt).

Maturity Model

Level 0: Reactive     -- Fighting fires, manual everything
Level 1: Defined      -- Documented processes, basic monitoring
Level 2: Measured     -- SLOs defined, error budgets tracked
Level 3: Automated    -- Self-healing, auto-scaling, GitOps
Level 4: Optimized    -- Proactive capacity, chaos engineering, continuous improvement

Compliance and regulatory frame

SRE work intersects with compliance frameworks and regulatory regimes. See 08-security for full cluster index. Quick map:

Cross-Project

  • Career Vault -- Outcome Engineering framework, [employer] accomplishments
    • Pillar 9 (Cost Optimization) connects to FinOps Practice
    • Pillar 4 (Incident Management) connects to Principle 7: Failures are System Data
    • Pillar 10 (Toil Reduction) connects to Principle 5: Automate or Justify
  • UncoverTechTalent -- interview training, SRE hiring signals
  • ThatFinOpsGuy -- FinOps consulting (deep overlap with Pillar 9)
  • Video Summaries -- tagged reference material

People as SRE substrate

Same discipline applies to the human substrate the system runs on. Home network = office network = work network — one physical surface, all of it under the same operating frame. The psychology / developmental-position / competency corpora are the people-substrate layer of this SRE practice. The Receipts Standard (~/.claude/projects/-Users-stefancoetzee/memory/feedback_receipts_standard.md) is the LLM-substrate application of the same principles: probe before assert, snapshot before change, receipts not vibes.

  • Bridge essay: Interview Training as Applied Clinical Psychology -- hiring/interviewing as the substrate-observability tooling; 11 explicit parallels between clinical-craft and interview-craft
  • Psychology cluster: Psychology MOC -- substrate atoms (ownership-psychology axis, real-emotional-maturity, nervous-system-regulation, narcissism, fawning, hyper-independence, caretaker-syndrome, learned-helplessness, locus-of-control, amygdala-as-smoke-detector)
  • Developmental-position cluster: Developmental Position and Distortion -- six-axis framework + Three-Layer Position Model (substrate / apparent / assigned) + Trauma Filter + Defense-Response Model
  • Interview cycle + competency MOCs: Interview Cycle · Competencies Home · PM Psychological Dimensions
  • Per-pillar people-substrate cross-links live inside each pillar README's "People-substrate cross-cluster" section (pillars 01, 02, 03, 04, 07, 08, 09, 10). Pillars 05 (IaC) and 06 (CI/CD) skipped -- no defensible mapping.

Recently updated in this space