Run #3387

{
  "ok": true,
  "failed": [],
  "hallucinated": [],
  "context_errors": []
}

{
  "pool_status": "see_db"
}

Tu es un agent de monitoring du pool agencecom.

Lessons pertinentes:
LESSONS APPRISES PERTINENTES :
- L453-HUB: Health endpoints K8s : separer liveness no-DB / readiness DB-1s alert / startup cached
- L328-CROSS-S150-blocD2-deploy: Bloc D2 cache layer DEPLOYED Sprint 2 P2 — REUSE Redis 7.0.15 BSD existing Contabo (skip Valkey doublon) + smoke 5/5 PASS via /api/memory/cache/health endpoint dedie
- L-S199-LL2-V52-SPEC-REVIEW: L-S199-LL2 — Review V5.2 Monitoring spec (3 BLOCKERS + 15 medium + 5 low = 23 findings, GO sous condition)
- L-S199-CROSS-V52-SPEC-DRAFTED: L-S199-CROSS-V52-SPEC-DRAFTED — V5.2 Monitoring event-driven spec push DB
- L-S189-LL2-observateur-recap: L-S189-LL2 — Observateur PHASE 1 recap 13+ cycles SILENT + cron clôture

Donnees pool actuelles:
see_db

CONTRAINTE STRICTE -- Tables autorisees UNIQUEMENT :
- agencecom.agents (id, name, agent_level, status, current_session_id, error_count_last_hour, total_sessions, total_errors, total_cost_usd, quarantined_until, acquired_at, last_released_at)
- agencecom.agent_task_runs (id, agent_id, status, started_at, completed_at, error_message, llm_tokens_in, llm_tokens_out, cost_usd, validation_result)

Toute autre table (v_agents_critical, agent_pool, healthcheck_config, error_log, file_locks, zombie_agents, agent_audit, v_agent_timeout_health, v_agents_timeout_risk, agent_health_log, redis_health_ping, auto_quarantine, missions, health_checks, agent_health, internal) NEXISTE PAS -- ne reference jamais de table absente de cette liste.

Identifie les agents zombies (busy >1h sans completion via started_at + status), error_count_last_hour > 5, ou drift quarantine (quarantined_until > now).
Donne 3 recommandations concretes en markdown bullet points avec actions SQL utilisant SEULEMENT les 2 tables autorisees ci-dessus.

**Identification des agents zombies / à risque**  
*Requêtes basées sur `see_db` (données pool actuelles)*

- Agents bloqués `busy` > 1 h sans complétion :  
  ```sql
  SELECT a.id, a.name, atr.id AS task_id, atr.started_at
  FROM agencecom.agents a
  JOIN agencecom.agent_task_runs atr ON a.id = atr.agent_id
  WHERE atr.status = 'running'
    AND atr.started_at < NOW() - INTERVAL '1 hour';
  ```

- Agents avec taux d’erreur élevé :  
  ```sql
  SELECT id, name, error_count_last_hour
  FROM agencecom.agents
  WHERE error_count_last_hour > 5;
  ```

- Agents en quarantaine persistante (`quarantined_until > now()`) :  
  ```sql
  SELECT id, name, quarantined_until, error_count_last_hour
  FROM agencecom.agents
  WHERE quarantined_until > NOW();
  ```

---

**3 Recommandations concrètes avec actions SQL**

- **1. Nettoyer les sessions zombies (busy > 1 h)**  
  Marquer les tâches en échec “zombie” et replacer les agents en quarantaine pour libérer le pool.  
  ```sql
  WITH zombie_tasks AS (
    SELECT atr.id AS task_id, atr.agent_id
    FROM agencecom.agent_task_runs atr
    WHERE atr.status = 'running'
      AND atr.started_at < NOW() - INTERVAL '1 hour'
  )
  UPDATE agencecom.agent_task_runs
  SET status = 'failed',
      completed_at = NOW(),
      error_message = 'Zombie detection: running over 1h'
  FROM zombie_tasks z
  WHERE agent_task_runs.id = z.task_id;
  
  UPDATE agencecom.agents
  SET status = 'quarantined',
      quarantined_until = NOW() + INTERVAL '15 minutes',
      current_session_id = NULL
  FROM zombie_tasks z
  WHERE agents.id = z.agent_id;
  ```

- **2. Isoler les agents instables (error_count_last_hour > 5)**  
  Placer en quarantaine tout agent non déjà isolé et dépassant le seuil, avec une fenêtre de quarantaine adaptée.  
  ```sql
  UPDATE agencecom.agents
  SET status = 'quarantined',
      quarantined_until = NOW() + INTERVAL '1 hour'
  WHERE error_count_last_hour > 5
    AND (status != 'quarantined' OR quarantined_until IS NULL OR quarantined_until <= NOW());
  -- Optionnel : réinitialiser le compteur après analyse manuelle
  -- UPDATE agencecom.agents SET error_count_last_hour = 0 WHERE id IN (...);
  ```

- **3. Libérer les quarantaines sans cause active**  
  Tout agent toujours marqué `quarantined_until > now()` mais avec `error_count_last_hour = 0` et sans tâche en cours peut être réintégré.  
  ```sql
  UPDATE agencecom.agents
  SET status = 'idle',
      quarantined_until = NULL
  WHERE quarantined_until IS NOT NULL
    AND quarantined_until > NOW()
    AND error_count_last_hour = 0
    AND NOT EXISTS (
      SELECT 1 FROM agencecom.agent_task_runs atr
      WHERE atr.agent_id = agents.id
        AND atr.status = 'running'
    );
  ```

*Ces actions utilisent strictement les tables autorisées : `agencecom.agents` et `agencecom.agent_task_runs`.*