topic Amazon RDS Metric Events: Migrating from Classic "built-in" Services Metrics to Cloud Services Metrics in Cloud platforms

Amazon RDS Metric Events: Migrating from Classic "built-in" Services Metrics to Cloud Services Metrics

ASE — Wed, 26 Jun 2024 13:07:36 GMT

Good afternoon.

We are in the process of migrating Amazon RDS from Classic Services to Cloud Services in Dynatrace.

Based on the documentation I read, the built-in RDS alerts shown below will not be available once we migrate.

Because of this, I'm working on re-creating these alerts using the new Cloud Services Metrics.

However, I'm stuck trying to duplicate the "Detect RDS running out of memory" alert.

How do I combine the "Freeable Memory" metric with the "Swap Usage" metric in my Metric Event?

If I'm reading this correctly, both conditions need to exist in order to trigger this alert.

Re: Amazon RDS Metric Events: Migrating from Classic "built-in" Services Metrics to Cloud Services Metrics

ASE — Fri, 28 Jun 2024 04:00:23 GMT

What about the RDS Free Storage Space % metric?

The % metric is not available in Amazon RDS Cloud Services. It's using the number of Bytes.

Classic Services

Cloud Services

I suspect we could use a Metric Expression to calculate the Free Storage Space %, but how is that possible if we are not ingesting an "allocated storage" metric from Amazon RDS?

Any recommendations on how to recreate this Classic alert using the Cloud metrics?

Re: Amazon RDS Metric Events: Migrating from Classic "built-in" Services Metrics to Cloud Services Metrics

ASE — Mon, 08 Jul 2024 15:16:18 GMT

Solution:

This is what we used to build the Cloud Services version of this Classic "built-in" RDS running out of memory alert:

ext:cloud.aws.rds.freeableMemory:splitBy("dt.entity.custom_device"):filter(series(avg,lt(100000000))):count:default(0, always) + ext:cloud.aws.rds.swapUsage:splitBy("dt.entity.custom_device"):filter(series(avg,gt(1000000000))):count:default(0, always)