Database-Driven Auto Scaling Pipeline

Scenario

Use this pipeline when your application’s compute demand is directly coupled to database load, such as during batch processing or traffic surges that cause RDS CPU or IOPS spikes. Instead of relying on ECS-level metrics alone, this setup creates a closed-loop response where RDS performance thresholds automatically drive ECS instance provisioning via EventBridge.

Integration steps

1. Enable RDS Metric Publishing: Following rds-monitor-performance, publish metrics to CloudMonitor at a 60s interval: aliyun rds ModifyDBInstanceMonitor --DBInstanceId <instance-id> --Period 60.
2. Create ESS Scaling Group & Rule: Per ess-configure-triggers, provision the group and define the scale-out action: aliyun ess CreateScalingGroup --ScalingGroupName db-driven-sg --MinSize 2 --MaxSize 10. Then: aliyun ess CreateScalingRule --ScalingGroupId <sg-id> --ScalingRuleName scale-out-on-db --AdjustmentType TotalCapacity --AdjustmentValue 2.
3. Configure EventBridge Event Source: As routed by eb-configure-streaming, register the RDS metric stream: aliyun eventbridge CreateEventBus --EventBusName rds-metrics-bus.
4. Define EventBridge Rule & Filter: Create a threshold rule matching the eb-monitor-alerts pattern:

``bash aliyun eventbridge PutRule --EventBusName rds-metrics-bus --RuleName rds-cpu-threshold \ --EventPattern '{"source":["acs.rds"],"type":["CloudMonitor:Metric"],"detail":{"metricName":"CPUUtilization","value":{">":80}}}' ``

5. Route to ESS Target: Attach the scaling rule ARN as the EventBridge target:

``bash aliyun eventbridge PutTargets --EventBusName rds-metrics-bus --RuleName rds-cpu-threshold \ --Targets '[{"Id":"1","Arn":"acs:ess:<region>:<account>:scalingRule/<rule-id>","Type":"acs.ess","Parameters":"{\"adjustmentType\":\"TotalCapacity\",\"adjustmentValue\":2}"}]' ``

6. Configure Monitoring Alerts: Enable delivery failure alerts per eb-monitor-alerts: aliyun eventbridge PutRuleAlert --RuleName rds-cpu-threshold --AlertThreshold 1 --Period 300.

Architecture

RDS continuously publishes performance metrics to CloudMonitor. EventBridge ingests these as structured events, applies JSONPath filters against defined thresholds, and routes matching payloads directly to the ESS scaling rule ARN. ESS executes the scale-out action, adding ECS instances to handle the increased query load. Alerts are generated if EventBridge fails to deliver or if scaling execution times out, closing the observability loop.

Prerequisites

• Active Alibaba Cloud account with AliyunRDSFullAccess, AliyunEventBridgeFullAccess, and AliyunESSFullAccess RAM roles.
• RDS instance with CloudMonitor integration enabled and metric reporting interval ≤ 60s.
• Pre-configured VPC and security group allowing new ECS instances to connect to the RDS endpoint.
• ESS scaling group with a valid launch template referencing the application image.

Common pitfalls

• Metric Lag: RDS publishes metrics at 60s intervals by default; scaling triggers may fire 1–2 minutes after the actual spike. Use Period: 30 in ModifyDBInstanceMonitor if supported.
• Event Pattern Mismatch: EventBridge requires exact JSON structure matches. Missing type or incorrect metricName casing (CPUUtilization vs cpu_utilization) silently drops events.
• Scaling Cooldown Conflicts: ESS default cooldown (300s) can block rapid scale-in/out cycles. Explicitly set Cooldown: 60 in CreateScalingGroup to match real-time DB load fluctuations.
• Permission Boundary: The EventBridge service role must explicitly allow ess:ExecuteScalingRule on the target ARN; otherwise, targets fail with AccessDenied.

Typical questions

• event-driven auto scaling
• 数据库驱动弹性伸缩
• stream RDS events to trigger scaling
• monitor database and auto scale
• RDS performance triggers scaling
• 数据库监控触发扩缩容
• event-based scaling pipeline
• real-time database event automation