I use the awscli tools to verify the size of my buckets. First the S3 call (all files without versioning, just an ordinary upload):
$ aws s3 ls s3://XXX --recursive --human-readable --summarize
(...)
Total Objects: 2
Total Size: 82.5 GiB
Then I used the CloudWatch metric:
$ aws cloudwatch get-metric-statistics --namespace "AWS/S3" --start-time 2015-08-29 --end-time 2015-08-31 --metric-name BucketSizeBytes --period 3600 --statistics Average --unit Bytes --dimensions Name=BucketName,Value=XXX Name=StorageType,Value=StandardStorage
{
"Datapoints": [
{
"Timestamp": "2015-08-30T00:00:00Z",
"Average": 141012339184.0,
"Unit": "Bytes"
}, (...)
Now I am wondering why CloudWatch shows a much higher number than the S3 call. I couldn't find any hint from the documentation and therefore would be grateful if anybody could help.
(Same numbers are shown in the management console; about 82.5 GiB on the S3 console and 141,012,339,184 bytes on the CloudWatch console.)