Advanced tagging: AWS makes it easy to break down invoices


AWS explains the process in a blog post. In particular, companies that have already come to the taste of the Amazon cloud infrastructure and use a corresponding number of services should facilitate the overview and assignment of invoice items to cost centers. The same goes for service providers who manage AWS resources for their customers.

The following AWS resources can now be tagged (tagged):

  • S3 bucket

  • EC2 instances

  • EBS volumes

  • Reserved instances

  • Spot Instance requests

  • VPN connections

  • Amazon RDS database instances

  • AWS cloud formation stacks

The user can assign up to ten tags per resource. For this he should initially create a meaningful system. One possibility, for example, is a key "department" with values ​​such as sales, marketing, development, etc., possibly equal to the existing accounting system. And if each department drives different AWS-based applications or stores a lot of data in S3, then a complementary application tag would also make sense whose values ​​list all the applications that a department accesses.

Afterwards, the already used resources have to be provided with the agreed tags - this is done via the management console, the service APIs, the command line or via Auto Scaling - as well as newly set up as well. Then you have to tell the system in the AWS portal via the function "Manage Cost Allocation Report", which tags are actually relevant for the cost allocation. AWS then considers this data. The expected bill is generated several times a day and the final monthly bill (which can also be automatically delivered to your S3 bucket) is generated within three days of the end of the month.

The generated Cost Allocation Report then contains an additional column for each day defined as relevant, in which the corresponding value may appear. Also, the cost of data transfer and requests are listed (they inherit the tags from the assigned resource).

You can then import this data into your own accounting system or further analyze and visualize it for reporting - for example, a pivot table could be used in which the data is aggregated along one or more dimensions. Amazon wants to expand the tagging system to more resources in the future and is also thinking about increasing the limit of ten tags per resource. In addition, the upfront costs for Reserved Instance purchases are taken into account. Further suggestions from users are also welcome anytime.