Labeling your belongings is something that we learn from an early age, so that if something gets lost, it can easily be identified and returned to its rightful owner.
The same concept is used in naming your files and documents, and tagging your cloud resources in AWS, Azure or Google Cloud. In enterprises where there may be thousands of virtual machines running from different cloud providers, tagging is particularly important so that resources can be easily defined, managed and charged back to the correct department or business unit, and companies can understand how to allocate their costs and the cost of production.
Tagging usually uses a naming convention based on instance names or purpose, for ease of identification and chargeback. For example, the tag: “prod_ws_app” refers to “production- web server- app name;” using “production” as the purpose. Using this convention, every tag that starts with “prod” will be tagged with the purpose: “production”.
In order for an enterprise to accurately assign its cloud resources according to purpose, owner or location, its tagging method needs to be consistent. Difficulties arise with the easy access that employees have to open up new cloud resources according to need. Different people use varying definitions for the same tag; for example, when defining the customer service team, one person may use the tag “Customer Service”, while another uses “Cust. Service” and still another uses the tag “CS”. These discrepancies are going to lead to problems in creating a single report for the cloud usage of the customer service team, and for billing them.
According to our data, based on over 400 million instances, 30% of instances use a tag that describes the purpose of the instance, and out of those, 35% are production related. An example of this is the tag: “environment”. For this one tag, our customers use more than 1500 versions for the same tag. The most common ones are shown in Fig 1. Below:
A further difficulty lies in the cloud resources that are not tagged at all. This could be due to inefficient work practices, or the fact that some resources are considered untaggable!
Why would a resource be untaggable?
Sometimes there are cost items / services that cannot easily be assigned to a particular department or business unit and are therefore hard to tag. Examples include the AWS line item: “support”, the shared resource “email system” and “data transfer”. How can these items be tagged? Do they relate to production or development, and which business unit would bear the costs?
So how can you solve this tagging issue?
There are tagging automation tools available, including Chef, Puppet and Ansible, that help to ensure that every new cloud resource that you provision is tagged. These tools help you make your tags consistent across the departments in your organization.
The Role of the Category Manager
Even if you use one of these tagging automation tools, it is prone to error, with multiple people across a single company using varying names for the same tag (see the different names for the “environment” tag above), and no blanket policy enforced across the company.
To remedy, this, Cloudyn’s cloud management solution includes the Category Manager which helps to provide a rule based policy for creating new tags that is clear and consistent, and incorporates every cloud resource, even if it is with a miscellaneous type of tag, so that every resource can be accounted for.
Fig 2: Cloudyn’s Category Manager
To summarize, tagging is a necessary part of cloud cost allocation and cost visibility, and without it, an enterprise will have many unassigned resources that cannot be accounted for, and difficulties defining production costs. Tagging and cost allocation enable IT to regain control of their company’s cloud usage, as well as keeping the folks in finance happy!
To sign up for a 20-day free Cloudyn trial, click here.