# =============================================================================
# COMPETENCY RECORD - Cedric Ancellin's demonstrated engineering skills and
# experience. Third-party technologies and vendors are named as the concrete
# skill surface. Professional background / track record.
# =============================================================================

# ============================================================
# RELIABILITY & OPERATIONS - COMPETENCY RECORD
# ============================================================
# Core Purpose:
#   Demonstrated experience running a production platform:
#   incident command, structured root-cause analysis, safe
#   data remediation at scale, error/alerting pipelines,
#   environment discipline, and idempotent operational tooling.
# ============================================================

summary: >
  Ran a production platform as an operational discipline -
  severity-classified incident response with formal postmortems,
  root-cause analysis across availability/integrity/exposure
  incidents, idempotent and reversible large-scale data
  remediation, an environment-gated error/alerting pipeline, and
  strict multi-environment operational hygiene.


# ==================================================================
# 1. INCIDENT COMMAND & POSTMORTEM DISCIPLINE
# ==================================================================

incident_management:
  description: >
    Established a formal incident-management practice with a
    consistent postmortem structure.
  severity_model:
    sev1: "Critical, customer-facing outage or confidentiality/privacy breach."
    sev2: "High, partial unavailability or data-integrity breakage beyond a single record."
    sev3: "Minor, internal/staging or limited-scope customer-facing issue."
    sev4: "Minimal, internal disruption, no customer impact."
  postmortem_fields:
    - "Severity, status lifecycle (draft → reviewed → actioned), duration, impact, detection."
    - "Root cause, contributing factors, resolution, and follow-up actions with ownership + due dates."
    - "Precise timelines, containment/eradication/recovery phases, and lessons learned."


# ==================================================================
# 2. ROOT-CAUSE ANALYSIS
# ==================================================================

rca:
  description: >
    Performed rigorous root-cause analysis across multiple failure
    classes, fixing the structural cause rather than the symptom.
  examples:
    data_integrity: >
      Diagnosed multiple writers disagreeing on a canonical
      identity field; resolved by consolidating to a single-writer
      model and a clarified identity model, not just repairing the
      corrupted rows.
    availability: >
      Diagnosed a worker-pool exhaustion caused by proxying
      long-lived connections through a request worker model;
      remediated the proxy path with follow-ups for a better
      concurrency model and rate limiting.
    data_exposure: >
      Diagnosed authenticated responses becoming edge-cache
      eligible during a mitigation; restored cache-bypass rules
      and reviewed application code for per-user non-cacheability.
    latent_bug: >
      Diagnosed unsafe relative include paths silently resolving
      to the wrong files; fixed the path model with follow-ups for
      absolute includes, environment isolation, and static analysis.


# ==================================================================
# 3. SAFE DATA REMEDIATION AT SCALE
# ==================================================================

remediation:
  description: >
    Built reusable, safe tooling for repairing production data at
    scale.
  principles:
    - "Dry-run by default; an explicit apply flag required to write."
    - "Idempotent and reversible via status flags rather than destructive edits."
    - "Re-derive from the source of truth rather than guessing corrected values."
    - "Independent verification by someone other than the engineer applying the fix."
  outcome_shape: "Generalized rebuild/rehome scripts repairing thousands of records across many properties in bounded compute windows, with per-property metrics."


# ==================================================================
# 4. ERROR HANDLING & ALERTING PIPELINE
# ==================================================================

alerting:
  description: >
    Built an environment-aware error and alerting pipeline.
  mechanisms:
    - "A single error-terminating path returning a structured error envelope, never an ad-hoc die/throw."
    - "Production/preprod errors fan out to chat + email; non-production suppresses unless debug is explicitly enabled."
    - "Per-datastore failure handlers with environment-gated verbosity."
    - "Cross-request alert throttling to prevent alert storms."
    - "A circuit breaker for cache-tier unavailability with a cooldown."


# ==================================================================
# 5. ENVIRONMENT & DEPLOYMENT DISCIPLINE
# ==================================================================

ops_hygiene:
  description: >
    Enforced strict operational hygiene across environments.
  practices:
    - "Production / UAT / staging / sandbox parity for security-relevant configuration."
    - "UAT validation before any live change; dry-run tooling before production writes."
    - "A git + publish deployment pipeline with a clean-working-tree guarantee; no direct edits to deployed paths."
    - "Sandboxing of throwaway/diagnostic scripts away from production code paths."
  cross_reference:
    $ref: "https://zc8.com/competencies/12_technical_leadership"