YACE Dashboard


1. Scope#

YACE 只能采集 RDS / Aurora PostgreSQL 的 CloudWatch 指标，不能替代 postgres_exporter。

当前这套 dashboard 按 YACE 输出里的 dimension_DBClusterIdentifier 过滤，更适合 Aurora PostgreSQL / cluster 维度的 metrics。

use YACE for:
    RDS / Aurora CPU / memory / connection / IO / cloud storage / replica lag

use postgres_exporter for:
    pg_stat_activity
    pg_stat_database
    pg_stat_statements
    locks / vacuum / table stats

2. YACE Config#

apiVersion: v1alpha1
sts-region: ap-northeast-1
discovery:
  jobs:
    - type: AWS/RDS
      regions: [ap-northeast-1]
      customTags:
        - key: environment
          value: uat
      period: 300
      length: 600
      nilToZero: true
      metrics:
        - name: CPUUtilization
          statistics: [Average]
        - name: FreeableMemory
          statistics: [Average]
        - name: DatabaseConnections
          statistics: [Average]
        - name: ReadLatency
          statistics: [Average, p95]
        - name: WriteLatency
          statistics: [Average, p95]
        - name: ReadIOPS
          statistics: [Average]
        - name: WriteIOPS
          statistics: [Average]
        - name: VolumeBytesUsed
          statistics: [Average]
          period: 300
          length: 900
        - name: AuroraReplicaLagMaximum
          statistics: [Maximum]
        - name: Deadlocks
          statistics: [Sum]

Aurora note:

当前 dashboard 按 cluster 维度工作，所以这里使用:
    VolumeBytesUsed
    AuroraReplicaLagMaximum

不是:
    FreeStorageSpace
    ReplicaLag

如果 cluster 没有 reader，Aurora replica lag panel 仍然可能是 no data。

`VolumeBytesUsed` 建议单独拉长查询窗口:

```yaml
- name: VolumeBytesUsed
  statistics: [Average]
  period: 300
  length: 900

原因:

这个 metric 不是每个很短窗口都稳定返回 datapoint。
length 比 period 更长，可以降低 YACE 因为短时取不到点而补成 0 的概率。

Dashboard 选择:

```text
aws-aurora-postgresql-yace-overview-dashboard.json:
    使用 $dimension_DBClusterIdentifier 过滤。
    适合单一环境，或者 cluster 名本身已经足够区分环境。

aws-aurora-postgresql-yace-env-overview-dashboard.json:
    使用 $environment + $dimension_DBClusterIdentifier 过滤。
    适合同一个 Prometheus / VictoriaMetrics 里同时存在 uat、prod 等多个环境的 Aurora / RDS cluster metrics。

3. Dashboard Rows#

Row Panels
Overview CPU, free memory, connections, read/write latency
Drilldown read/write IOPS, deadlocks, Aurora replica lag
Capacity storage used, connections trend, IOPS trend
Debug latency split and deadlock trend

Trend note:

带 Trend 的 panel 不是原始单点值。

当前 Aurora PostgreSQL dashboard 里:
    Connections Trend
        最近 1 小时 rolling average

    IOPS Trend
        最近 1 小时 rolling average

    Deadlock Trend
        最近 1 小时 rolling sum

4. Verify#

{__name__=~"aws_rds_.*"}
CloudWatch Dashboard Forward