Cloudwatch
要启用ASG监控采集
图表 | 指标 | 截图里选哪一个 | Statistic | Period |
01-ALB-TG-Request | RequestCount | ApplicationELB > Per AppELB Metrics | Sum | 1 minute |
RequestCountPerTarget | ApplicationELB > Per AppELB, per TG Metrics | Sum | 1 minute | |
02-ALB-TG-5XX | HTTPCode_ELB_5XX_Count | ApplicationELB > Per AppELB Metrics | Sum | 1 minute |
HTTPCode_Target_5XX_Count | ApplicationELB > Per AppELB, per TG Metrics | Sum | 1 minute | |
HTTPCode_ELB_503_Count(有就加) | ApplicationELB > Per AppELB Metrics | Sum | 1 minute | |
03-ALB-TG-Latency | TargetResponseTime | ApplicationELB > Per AppELB, per TG Metrics | Average | 1 minute |
TargetResponseTime | ApplicationELB > Per AppELB, per TG Metrics | p95 | 1 minute | |
04-TG-Health | HealthyHostCount | ApplicationELB > Per AppELB, per TG Metrics | Minimum | 1 minute |
UnHealthyHostCount | ApplicationELB > Per AppELB, per TG Metrics | Maximum | 1 minute | |
05-ASG-Capacity | GroupDesiredCapacity | 不是 ApplicationELB;返回全部 → Auto Scaling →按 Auto Scaling Group | Average | 1 minute |
GroupInServiceInstances | Auto Scaling →按 Auto Scaling Group | Average | 1 minute | |
GroupPendingInstances | Auto Scaling →按 Auto Scaling Group | Maximum | 1 minute |
告警
告警名称 | CloudWatch进入路径 | 资源选择 | 指标 Metric | Statistic | Period | 告警条件 |
CC-ALARM-UnhealthyTarget | ApplicationELB → Per AppELB, per TG Metrics | CC-ALB + CC-TG | UnHealthyHostCount | Maximum | 1 minute | >= 1 |
CC-ALARM-Target5XX | ApplicationELB → Per AppELB, per TG Metrics | CC-ALB + CC-TG | HTTPCode_Target_5XX_Count | Sum | 1 minute | >= 1 |
CC-ALARM-HighLatency | ApplicationELB → Per AppELB, per TG Metrics | CC-ALB + CC-TG | TargetResponseTime | p95 | 1 minute | >= 2 秒 |
CC-ALARM-ELB5XX(可选) | ApplicationELB → Per AppELB Metrics | CC-ALB | HTTPCode_ELB_5XX_Count | Sum | 1 minute | >= 1 |
CC-ALARM-RDS-HighCPU | RDS → DBInstanceIdentifier | cc-rds | CPUUtilization | Average | 1 minute | >= 65% |
ASG
策略 | 用途 | 指标 | 条件 | 动作 |
现有 Target Tracking | 正常自动扩缩 | RequestCountPerTarget | Target=200 | 自动加减实例 |
新增 Step Scaling | 突发流量快速扩容 | RequestCountPerTarget | >=400 | +1 |
>=600 | +2 | |||
>=900 | +3 |
先创建告警
CC-ALARM-BurstTraffic
Per AppELB, per TG Metrics --> RequestCountPerTarget
Statistic = Sum
Period = 1 minute