S3 Glacier storage with Glaciers3ObjectOverhead

Question

Hi,

I had a S3 bucket where most of my storage was in StandardStorage class and after 30 days objects were moved to Standard-IA class.

I have decided to change this behavior and to move all objects to Glacier on the day 1. On Metrics tab, I can now see that there are also objects marked as GlacierObjectOverhead & GlacierS3ObjectOverhead. From what I understand from the documentation, those are used to store metadata (such as name) about each "base" object. Which is weird is I have only 4Gb of objects stored but almost 60 Gb of GlacierObjectOverhead objects. So my question is : wouldn't it be more expensive to use Glacier ? Why do we have this weird behavior where metadata takes more space than the object itself (4 vs 60 Gb) ?

Thank you

Didier_Durand · Answer

Hi,

You 4 GB are probably made of many small objects rather than big ones.
Then the overhead is computed this way:

```
GlacierObjectOverhead – For each archived object, S3 Glacier adds 32 KB of storage for index and related 
metadata. This extra data is necessary to identify and restore your object. You are charged S3 Glacier 
Flexible Retrieval rates for this additional storage.

GlacierS3ObjectOverhead – For each object archived to S3 Glacier Flexible Retrieval, Amazon S3 uses 8 KB 
of storage for the name of the object and other metadata. You are charged S3 Standard rates for this 
additional storage.
```

Source: https://docs.aws.amazon.com/AmazonS3/latest/userguide/metrics-dimensions.html

So, compute with you number of objects (that you don't mention) to see if the 60GB in Glacier look correct. Does it explain the "weird behavior" as you describe it?

Glacier Deep Archive  (if you use this option) is approx 20x cheaper than S3 Standard. So, you would still have a (small) cost benefit to go to Glacier. See https://aws.amazon.com/s3/pricing/

Best,

Didier

S3 Glacier storage with Glaciers3ObjectOverhead

相关内容