# 参考 Pigsty v4.5.0 etcd 指标与阈值,适配原生 up 和现有标签;无 Pigsty UI 依赖。 apiVersion: operator.victoriametrics.com/v1beta1 kind: VMRule metadata: name: shared-etcd namespace: monitoring spec: groups: - name: shared-etcd rules: - alert: SharedEtcdMemberUnavailable expr: up{job="shared-etcd"} == 0 for: 2m labels: severity: warning service: shared-etcd annotations: summary: 共享 etcd 成员 {{ $labels.member }} 无法采集 - alert: SharedEtcdMonitoringQuorumUnavailable expr: (sum(up{job="shared-etcd"}) or vector(0)) < 2 for: 2m labels: severity: critical service: shared-etcd annotations: summary: 共享 etcd 不足两个成员可采集,请用 endpoint health 核实 quorum - alert: SharedEtcdNoLeader expr: etcd_server_has_leader{job="shared-etcd"} == 0 for: 1m labels: severity: critical service: shared-etcd annotations: summary: 共享 etcd 成员 {{ $labels.member }} 无法确认 leader - alert: SharedEtcdQuotaHigh expr: etcd_mvcc_db_total_size_in_bytes{job="shared-etcd"} / etcd_server_quota_backend_bytes{job="shared-etcd"} > 0.8 for: 10m labels: severity: warning service: shared-etcd annotations: summary: 共享 etcd 后端容量超过 80% - alert: SharedEtcdSlowFsync expr: histogram_quantile(0.95, rate(etcd_disk_wal_fsync_duration_seconds_bucket{job="shared-etcd"}[5m])) > 0.05 for: 10m labels: severity: warning service: shared-etcd annotations: summary: 共享 etcd {{ $labels.member }} WAL fsync p95 超过 50ms - alert: SharedEtcdFrequentElections expr: increase(etcd_server_leader_changes_seen_total{job="shared-etcd"}[15m]) > 3 for: 5m labels: severity: warning service: shared-etcd annotations: summary: 共享 etcd 15 分钟内多次选举