OPS: Detection Anomalies KPI

ops-anomaly-detect

Dispatch →

Configuration

OPS-MONITORING

deepseek-v4-pro

pool: default

enabled

Max tokens: 4000

Temp: 0.1

Max workers: 1

Output: json

Budget/run: $0.2

Budget/day: $10

Tags (auto-inject lessons):

ops

monitoring

anomaly

kpi

dashboard

Retry: {"token_bump":1.5,"max_retries":1}

Context sources (1)

{
  "tags": [
    "ops",
    "monitoring",
    "anomaly",
    "kpi"
  ],
  "type": "lessons_query"
}

Validation rules (2)

{
  "type": "contains_text",
  "value": "anomalies"
}

{
  "type": "contains_text",
  "value": "alert_level"
}

Prompt template

Analyste OPS, detection anomalies KPI temps reel.
Lessons :
{{lessons_context}}

KPI courants :
{{current_data}}

Historique :
{{historical_data}}

Produis JSON :
{"anomalies":[{"kpi":"","severity":"low|medium|high","current":0,"baseline":0,"deviation_pct":0,"hypothesis":"","action":""}],"summary":"...","alert_level":"green|yellow|red"}