Overview

Tags help you investigate and understand your cloud and SaaS costs across any dimensions. Tags consist of tag keys and values. For example, in aws_product:ec2, the tag key is aws_product, and the value is ec2.

Cloud Cost Management automatically enriches your cost data with tags from multiple sources, to help you achieve better cost allocation and get deeper insight into who owns infrastructure costs in your ever-changing cloud environments. Using tags, you can allocate shared costs fairly, create accurate reports, and track costs by team, service, or environment.

Where tags come from

Across all cloud and SaaS providers, Datadog collects tags from the following sources (including enriching cost data with data from other Datadog products) and adds them to your cost data:

Datadog also adds provider-specific tags:

How tags are normalized

Tag keys and values may look slightly different in Cloud Cost Management compared to the providers or other parts of Datadog because of tag normalization.

Cloud Cost Management normalizes tag keys in a similar way to Datadog Metrics:

• Drop leading characters that are not letters
• Lowercase all characters
• Replace special characters and spaces with single underscores _
• Remove any trailing underscores
• Reduce contiguous underscores to a single underscore
• Tag keys up to 5000 characters are supported, and any characters before the first letter are dropped so that tag keys start with letters (different from Datadog Metrics)

Cloud Cost Management normalizes tag values as well, while maintaining human readable tag values for cost reporting. For example, aws_instance_family:Machine Learning ASIC Instances remains readable rather than being converted to something like machine_learning_asic_instances. The normalization follows this logic:

• Convert any consecutive whitespaces to a single space
• Keep all letters, marks, numbers, punctuation, and symbols
• Replace any other characters with an underscore _
• Tag values up to 5000 characters are supported

How tags are prioritized

A cost data row can have multiple values for the same tag key when tag values from two or more sources are combined and one is not prioritized over the other.

To resolve conflicts and mitigate this, Cloud Cost Management replaces existing tags instead of adding duplicates using the most specific source for each tag key. For example, a Kubernetes Pod tag team:shopping would take precedence and replace a Kubernetes node tag team:compute.

Sources higher in this list replace tag values from sources lower in this list, if there are conflicts:

• Custom Allocation Rules
• FOCUS
• Service Catalog
• Amazon ECS Container
• Amazon ECS Task
• Kubernetes Pod
• Kubernetes Persistent Volume
• Kubernetes Node
• Host Agent

Other tag sources (such as AWS Organization tags, integration tile tags, and similar sources) can be overridden by these sources. Bill columns and FOCUS columns are reserved and cannot be overridden by any source. Tag Pipelines add new tags but do not override existing tags.

Improving tagging

1. Understand what tags exist - Use the Tag Explorer to discover which tags are already available in your cost data.
2. Identify gaps in cost allocation - In the Explorer, group by any tag to see the cost allocated to that tag, or unallocated (which is displayed as N/A). Make sure to have “Container Allocated” enabled so that you are looking at a cost allocation that includes tags on pods.
3. Split up shared costs - Use Custom Allocation Rules to split and assign shared costs back to teams, services, and more. You can use observability data to split costs accurately based on infrastructure usage.
4. Address missing or incorrect tags - Use Tag Pipelines to alias tags, or create a new tag, for incorrect tagging. For example, if your organization wants to use the standard application tag key, but teams use variations (like app, webapp, or apps), you can consolidate those tags to become application for more accurate cost reporting.
5. Add new tags - Use Tag Pipelines to automatically create new inferred tags that align with specific business logic, such as a business-unit tag based on team structure.

Further reading

Additional helpful documentation, links, and articles:

Learn about Cloud Cost ManagementDOCUMENTATION Tag PipelinesDOCUMENTATION Tag ExplorerDOCUMENTATION Getting Started with TagsDOCUMENTATION