Runbook: RDSIOPSUtilization Alert

Alert Details

  • Alert Name: RDSIOPSUtilization
  • Expression: (max by (cluster) (rds_read_iops_average{%(clusterSelector)s}) + max by (cluster) (rds_write_iops_average{%(clusterSelector)s})) / max by (cluster) (rds_instance_max_iops_average{%(clusterSelector)s}) * 100 > 80

Description

Alert is triggered when RDS instance is close to its max allocated IOPS.

Possible Causes

  • high load

Troubleshooting Steps

1. Open the RDS server main dashboard

2. Open RDS Performance insights to identify IOPS-intensive queries

3. Kill SQL queries that generate intensive IOPS

4. If specific to an application, report the issue to the application owner to improve the query

5. Increase provisioned IOPS if possible (gp3 or io1 storage class). Be aware that some instances have IOPS limits.