I'm trying to understand how Datadog measures the total storage space for Aurora clusters. I can only find the FreeStorageSpace metric in CloudWatch, but I'm looking for something that provides a total storage metric. Any insights on how this works?
4 Answers
Isn't the storage size for Aurora clusters automatically growing? It seems almost unnecessary to monitor it unless you're nearing the maximum capacity, like 256 TB. But I guess if you want to manage costs, checking how much storage you're actually using could help.
This confused me in the beginning too. Datadog's total_storage_space is mostly inferred since Aurora doesn't pre-provision storage and there isn't a hard cap metric in CloudWatch. You can look at VolumeBytesUsed for real usage and refer to the docs for the cluster's max size (like 128 TiB for MySQL). Essentially, Datadog is doing the math with what's available. And by the way, CubeAPM is gaining traction for allowing custom metrics, which is pretty cool for needs like this.
I can't check at the moment, but I believe there's a metric called "VolumeBytes" in the RDS console. It should show you the used space for your Aurora cluster.
I'm not entirely sure how Datadog does it, but I would use the `VolumeBytesUsed` CloudWatch metric for Aurora. This shows how much storage your Aurora DB cluster is using, which directly affects your costs. Just remember, it doesn't account for some internal storage allocations that won’t impact your billing. For better monitoring, watch `AuroraVolumeBytesLeftTotal` too, especially if you're worried about running out of space!
That's true! It's better to keep an eye on it, especially given the costs associated with Aurora storage.
That sounds interesting! Custom metrics could really help with specific monitoring needs.