memoryweave/deploy/prometheus-alerts.yml

86 lines
3.4 KiB
YAML
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# MemoryWeave (织忆) Prometheus 告警规则
# 位置: /etc/prometheus/rules/zhiyi.yml
groups:
- name: zhiyi_alerts
rules:
# ─── 蒸馏积压 ─────────────────────────
- alert: ZhiyiDistillBacklog
expr: zhiyi_distill_queue_depth > 50
for: 10m
labels:
severity: warning
annotations:
summary: "织忆蒸馏队列积压 ({{ $value }} > 50)"
description: "蒸馏队列深度超过 50可能出现处理瓶颈"
# ─── 冲突积压 ─────────────────────────
- alert: ZhiyiConflictsPending
expr: zhiyi_conflicts_pending > 5
for: 15m
labels:
severity: warning
annotations:
summary: "织忆待解决冲突 ({{ $value }} > 5)"
description: "有 {{ $value }} 个冲突等待裁决,需要牧尘介入"
# ─── Recall 延迟异常 ───────────────────
- alert: ZhiyiRecallLatencyHigh
expr: histogram_quantile(0.99, rate(zhiyi_request_duration_seconds_bucket{endpoint="recall"}[5m])) > 2
for: 5m
labels:
severity: warning
annotations:
summary: "织忆 recall P99 延迟 > 2s ({{ $value }}s)"
description: "recall 接口变慢,检查 BGE/reranker 服务状态"
# ─── 服务宕机 ─────────────────────────
- alert: ZhiyiServiceDown
expr: up{job="zhiyid"} == 0
for: 1m
labels:
severity: critical
annotations:
summary: "织忆服务宕机"
description: "zhiyid 进程在 {{ $labels.instance }} 上停止响应"
# ─── LLM 调用额度接近上限 ─────────────
- alert: ZhiyiLLMBudgetNearLimit
expr: zhiyi_distill_llm_calls_today / 50 > 0.85
for: 30m
labels:
severity: warning
annotations:
summary: "织忆 LLM 调用接近日限 ({{ $value | humanizePercentage }})"
description: "今日已用 {{ $value }}% 的蒸馏额度,将降级为硬规则提取"
# ─── 深度整合失败 ─────────────────────
- alert: ZhiyiConsolidationFailing
expr: rate(zhiyi_consolidate_failures_total[1h]) > 0.05
for: 30m
labels:
severity: critical
annotations:
summary: "织忆深度整合连续失败"
description: "zhiyi-consolidate 连续 3 次失败,需手动检查 Rust sidecar"
# ─── 内存使用过高 ─────────────────────
- alert: ZhiyiMemoryHigh
expr: zhiyi_memory_rss_bytes > 200e6
for: 10m
labels:
severity: warning
annotations:
summary: "织忆内存使用 > 200MB ({{ $value | humanize }}B)"
description: "RSS 内存超过预期,可能存在内存泄漏"
# ─── SQLite 数据库过大 ────────────────
- alert: ZhiyiSQLiteDBSize
expr: zhiyi_sqlite_db_size_bytes > 1e9
for: 1h
labels:
severity: warning
annotations:
summary: "织忆 SQLite 数据库 > 1GB ({{ $value | humanize }}B)"
description: "数据库持续增长,建议运行深度整合修剪"