{
  "apiVersion": "v1",
  "kind": "PostMortem",
  "metadata": {
    "name": "PM-20260318-001",
    "incident": "INC-20260318-003",
    "createdAt": "2026-03-18T23:00:00Z",
    "author": "ai-generated",
    "reviewedBy": "carlos.silva@empresa.com"
  },
  "report": {
    "title": "OOMKill recorrente no checkout-service",
    "severity": "critical",
    "duration": "45m",
    "detectedAt": "2026-03-18T22:00:00Z",
    "resolvedAt": "2026-03-18T22:45:00Z",
    "summary": "O checkout-service sofreu multiplos OOMKills durante 45 minutos devido a um memory leak no handler de webhook de pagamento. O incidente afetou 3.2% das requests de checkout e foi resolvido automaticamente pelo operator apos aprovacao.",
    "impact": {
      "usersAffected": 1250,
      "requestsAffected": "3.2%",
      "revenueImpact": "estimado R$ 15.000",
      "slosBreached": ["slo-checkout-error-rate"]
    },
    "rootCause": {
      "description": "Memory leak no handler de webhook de pagamento. O handler nao fechava o body do HTTP response em caso de erro de validacao, causando acumulo de buffers na heap.",
      "category": "code-bug",
      "confidence": 0.91
    },
    "timeline": [
      {
        "time": "2026-03-18T22:00:00Z",
        "event": "Primeiro OOMKill detectado no pod checkout-service-5f8d7c6b4-k9m2n"
      },
      {
        "time": "2026-03-18T22:02:00Z",
        "event": "Decision engine inicia analise automatica"
      },
      {
        "time": "2026-03-18T22:03:30Z",
        "event": "Root cause identificado: memory leak no webhook handler"
      },
      {
        "time": "2026-03-18T22:05:00Z",
        "event": "Runbook 'oomkill-standard' selecionado, aprovacao solicitada"
      },
      {
        "time": "2026-03-18T22:08:00Z",
        "event": "Aprovacao concedida por carlos.silva@empresa.com"
      },
      {
        "time": "2026-03-18T22:10:00Z",
        "event": "Memory limit aumentado para 1Gi, pods reiniciando"
      },
      {
        "time": "2026-03-18T22:45:00Z",
        "event": "Incidente resolvido — pods estaveis por 30 minutos"
      }
    ],
    "remediationApplied": {
      "runbook": "runbook-oomkill-standard",
      "steps": [
        "Diagnostico de memoria via kubectl top",
        "Patch de memory limit de 512Mi para 1Gi",
        "Verificacao de estabilidade por 30 minutos"
      ],
      "result": "success"
    },
    "actionItems": [
      {
        "priority": "P1",
        "description": "Corrigir memory leak no handler de webhook — fechar response body em todos os paths de erro",
        "assignee": "backend-team",
        "status": "open",
        "dueDate": "2026-03-22"
      },
      {
        "priority": "P2",
        "description": "Adicionar metricas de heap usage no checkout-service para deteccao precoce",
        "assignee": "platform-team",
        "status": "open",
        "dueDate": "2026-03-25"
      },
      {
        "priority": "P3",
        "description": "Revisar memory limits de todos os servicos do namespace production",
        "assignee": "sre-team",
        "status": "open",
        "dueDate": "2026-03-30"
      }
    ],
    "lessonsLearned": [
      "Add automated linting to verify response body closure in HTTP handlers",
      "Memory limits should be reviewed quarterly based on actual usage metrics",
      "OOMKill runbook worked well, but approval took 3 minutes — consider auto-approve for riskScore < 50"
    ],
    "metricSnapshots": [
      {"name": "cpu_usage", "value": "0.45", "timestamp": "2026-03-18T21:55:00Z", "phase": "before"},
      {"name": "memory_usage", "value": "498000000", "timestamp": "2026-03-18T22:00:00Z", "phase": "during"},
      {"name": "memory_usage", "value": "312000000", "timestamp": "2026-03-18T22:30:00Z", "phase": "after"}
    ],
    "blastRadius": [
      {"resource": {"kind": "Service", "name": "checkout-svc", "namespace": "production"}, "impact": "5xx responses during pod restarts", "severity": "high"}
    ],
    "gitCorrelation": {
      "commitSHA": "a1b2c3d4",
      "commitMessage": "feat: add webhook handler for payment notifications",
      "author": "dev@company.com",
      "timestamp": "2026-03-18T19:30:00Z",
      "confidence": 0.82,
      "filesChanged": ["internal/webhook/handler.go", "internal/webhook/handler_test.go"]
    },
    "trending": {
      "occurrenceCount": 3,
      "windowDays": 30,
      "relatedPostMortems": ["PM-20260305-001", "PM-20260312-002"],
      "pattern": "Recurring oom_kill on Deployment/checkout-service (3 occurrences in 30 days)"
    },
    "gitOpsContext": "Helm release 'checkout' chart=checkout version=2.3.1 status=deployed revision=42",
    "logAnalysisSummary": "1 Go panic stack trace; 12 critical error patterns (resource/connectivity); Primary exception: panic: runtime error: invalid memory address",
    "cascadeChain": ["production/checkout-service(root_cause)", "production/api-gateway(victim)"],
    "feedback": {
      "overrideRootCause": "",
      "remediationAccuracy": 4,
      "comments": "Good analysis. Consider suggesting AdjustResources before restart next time.",
      "providedBy": "sre@company.com",
      "providedAt": "2026-03-19T09:00:00Z"
    }
  }
}