Alert on RDS balance metrics

[msvticket]

Objective

Alert on balance metrics getting low. The related CloudWatch metrics are:

• BurstBalance
• EBSIOBalance%
• EBSByteBalance%

Why

When any of these balances reach 0 the performance of an RDS server can deteriorate quite drastically, so getting an alert of the risk for this would be good.

How

• Start fetching these metrics in prometheus-rds-exporter
• Implement alerts in the prometheus-rds-alerts chart.

References

• https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/rds-metrics.html
• https://aws.amazon.com/blogs/compute/improving-application-performance-and-reducing-costs-with-amazon-ebs-optimized-instance-burst-capability/
• https://joelngwt.github.io/2021/08/27/understanding-aws-rds-burst-balance.html