{ "title": "Пайплайны пост-обработки", "description": "Требует профилей compose transcribe/llm (см. .env.example, COMPOSE_PROFILES) — по умолчанию (media,monitoring) большинство панелей будет показывать «No data», это ожидаемо, не баг.", "uid": "vidconf-pipelines", "editable": false, "timezone": "browser", "schemaVersion": 39, "version": 1, "time": { "from": "now-6h", "to": "now" }, "refresh": "30s", "tags": ["vidconf", "pipeline"], "panels": [ { "id": 1, "title": "Сеансы по статусу пайплайна", "description": "conference_sessions.pipeline_status (recording→transcribing→summarizing→notified|failed), метрика vidconf_pipeline_sessions.", "type": "timeseries", "gridPos": { "h": 8, "w": 12, "x": 0, "y": 0 }, "datasource": { "type": "prometheus", "uid": "prometheus" }, "fieldConfig": { "defaults": { "custom": { "drawStyle": "line", "fillOpacity": 10, "stacking": { "mode": "normal" } } }, "overrides": [] }, "targets": [ { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "vidconf_pipeline_sessions", "legendFormat": "{{status}}", "refId": "A" } ] }, { "id": 2, "title": "Глубина очередей Celery", "description": "vidconf_celery_queue_depth{queue=...} — transcription/summarize/notify/celery (redis LLEN). Порог алерта QueueGrowing см. deploy/monitoring/alerts.yml.", "type": "timeseries", "gridPos": { "h": 8, "w": 12, "x": 12, "y": 0 }, "datasource": { "type": "prometheus", "uid": "prometheus" }, "fieldConfig": { "defaults": { "custom": { "drawStyle": "line", "fillOpacity": 10 } }, "overrides": [] }, "targets": [ { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "vidconf_celery_queue_depth", "legendFormat": "{{queue}}", "refId": "A" } ] }, { "id": 3, "title": "Латентность API (p50/p95/p99 по маршрутам)", "description": "vidconf_http_request_duration_seconds — histogram латентности HTTP по шаблону маршрута.", "type": "timeseries", "gridPos": { "h": 8, "w": 12, "x": 0, "y": 8 }, "datasource": { "type": "prometheus", "uid": "prometheus" }, "fieldConfig": { "defaults": { "unit": "s", "custom": { "drawStyle": "line", "fillOpacity": 5 } }, "overrides": [] }, "targets": [ { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "histogram_quantile(0.50, sum(rate(vidconf_http_request_duration_seconds_bucket[5m])) by (le, path))", "legendFormat": "p50 {{path}}", "refId": "A" }, { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "histogram_quantile(0.95, sum(rate(vidconf_http_request_duration_seconds_bucket[5m])) by (le, path))", "legendFormat": "p95 {{path}}", "refId": "B" }, { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "histogram_quantile(0.99, sum(rate(vidconf_http_request_duration_seconds_bucket[5m])) by (le, path))", "legendFormat": "p99 {{path}}", "refId": "C" } ] }, { "id": 4, "title": "Длительность шагов пайплайна (p95)", "description": "Ожидает метрику vidconf_pipeline_step_duration_seconds (histogram, label step) — пока не реализована (backend/api/metrics.py содержит только латентность HTTP и gauge'и статусов/очередей). Панель — заготовка под будущую инструментацию шагов transcribing/summarizing/notified; до её появления показывает «No data».", "type": "timeseries", "gridPos": { "h": 8, "w": 12, "x": 12, "y": 8 }, "datasource": { "type": "prometheus", "uid": "prometheus" }, "fieldConfig": { "defaults": { "unit": "s", "custom": { "drawStyle": "line", "fillOpacity": 5 } }, "overrides": [] }, "targets": [ { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "histogram_quantile(0.95, sum(rate(vidconf_pipeline_step_duration_seconds_bucket[15m])) by (le, step))", "legendFormat": "p95 {{step}}", "refId": "A" } ] }, { "id": 5, "title": "LLM-сервер доступен (job=llm)", "description": "up{job=\"llm\"} — 1, если Prometheus успешно скрейпит llama.cpp (алерт LlmDown, deploy/monitoring/alerts.yml). Job llm и алерт LlmDown закомментированы по умолчанию (см. prometheus.yml/alerts.yml) — раскомментируйте оба на инсталляциях с профилем llm/llm-gpu (пресеты 3–5), иначе эта панель показывает «No data».", "type": "stat", "gridPos": { "h": 4, "w": 6, "x": 0, "y": 16 }, "datasource": { "type": "prometheus", "uid": "prometheus" }, "fieldConfig": { "defaults": { "mappings": [ { "type": "value", "options": { "0": { "text": "DOWN", "color": "red" }, "1": { "text": "UP", "color": "green" } } } ], "thresholds": { "mode": "absolute", "steps": [{ "color": "red", "value": null }, { "color": "green", "value": 1 }] } }, "overrides": [] }, "targets": [ { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "up{job=\"llm\"}", "refId": "A" } ] }, { "id": 6, "title": "Сеансы failed (текущее число)", "description": "vidconf_pipeline_sessions{status=\"failed\"} — алерт PipelineFailed срабатывает на росте за 15 минут, здесь — снимок текущего значения.", "type": "stat", "gridPos": { "h": 4, "w": 6, "x": 6, "y": 16 }, "datasource": { "type": "prometheus", "uid": "prometheus" }, "fieldConfig": { "defaults": { "thresholds": { "mode": "absolute", "steps": [{ "color": "green", "value": null }, { "color": "red", "value": 1 }] } }, "overrides": [] }, "targets": [ { "datasource": { "type": "prometheus", "uid": "prometheus" }, "expr": "vidconf_pipeline_sessions{status=\"failed\"}", "refId": "A" } ] } ] }