--- title: Tags description: Learn how tags are sourced, enriched, and managed in Cloud Cost Management. breadcrumbs: Docs > Cloud Cost Management > Tags --- # Tags ## Overview{% #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{% #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: | Source | What tags are collected | Description | | ----------------------- | ---------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | All Providers | Bill columns | Such as AWS Cost and Usage Report (CUR) columns, Google Billing Export columns, and so on | | Datadog Enrichment | Host Agent | Tags added to host metadata by the Datadog Agent running on the host | | Datadog Enrichment | Software Catalog | Tags associated with this service in the APM Service Catalog | | Datadog Enrichment | Integration Tiles | Tags added to the Datadog integration tile for a specific cloud account. Integration tile tags apply to all costs in that account. Requires enabling the provider integration for each account | | Datadog Enrichment | Data Observability | Tags from Datadog Data Observability, powering BigQuery cost allocation. Requires enabling BigQuery monitoring | | Datadog Enrichment | Cloud Network Monitoring | Source and destination dimensions from [Datadog Cloud Network Monitoring](https://docs.datadoghq.com/network_monitoring/cloud_network_monitoring/). Requires enabling Cloud Network Monitoring in the Datadog Agent. See [data transfer cost allocation](https://docs.datadoghq.com/cloud_cost_management/allocation/container_cost_allocation/?tab=aws#data-transfer) for more details | | Kubernetes Enrichment | Kubernetes Node | User-defined tags found on Kubernetes nodes monitored with Datadog | | Kubernetes Enrichment | Kubernetes Pod | User-defined tags found on Kubernetes pods monitored with Datadog | | Kubernetes Enrichment | Kubernetes Persistent Volume | User-defined tags found on Persistent Volumes in Kubernetes clusters monitored with Datadog | | Kubernetes Enrichment | Kubernetes Persistent Volume Claim | User-defined tags found on Persistent Volume Claims in Kubernetes clusters monitored with Datadog | | Cloud Cost Management | Cloud Cost Aliases | Tags derived from provider cost data to simplify the cost data model, such as `aws_product` (an alias of `lineItem/ProductCode`). Additional tags that exist on both cost data and integration metrics are added, allowing cost data and usage data to be combined in Custom Allocation Rules, dashboards, and notebooks | | Cloud Cost Management | Cloud Cost Allocation | Tags created during [cost allocation](https://docs.datadoghq.com/cloud_cost_management/allocation/container_cost_allocation/) that specify the split of shared resources, such as `allocated_spend_type` | | Cloud Cost Management | FOCUS | Provider-agnostic tags compliant with [FOCUS](https://focus.finops.org/), an open specification that normalizes cost and usage datasets across cloud vendors | | Tag Pipelines | Rules defined by user | Tags created by applying your Tag Pipelines to cost data | | Custom Allocation Rules | Rules defined by user | Tags created by applying your Custom Allocation Rules to cost data (does not apply to SaaS costs) | Datadog also adds provider-specific tags: | Provider | What tags are collected | Description | | ------------ | ---------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | AWS | Cost Allocation Tags | Any tag defined by the user in [AWS Cost Allocation](https://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/cost-alloc-tags.html) tags that show up in the AWS CUR | | AWS | AWS Resource Groups Tagging API | User-defined tags on a cloud resource in AWS, pulled by Cloud Cost Management using the [Groups Tagging API](https://docs.aws.amazon.com/resourcegroupstagging/latest/APIReference/overview.html) | | AWS | AWS Organizational Units | User-defined tags on AWS organizational units using [AWS Organizations](https://docs.aws.amazon.com/organizations/latest/userguide/orgs_getting-started_concepts.html) | | AWS | AWS Organizations - Accounts | User-defined tags on an AWS account using [AWS Organizations](https://docs.aws.amazon.com/organizations/latest/userguide/orgs_getting-started_concepts.html) | | Amazon ECS | Amazon ECS Task | User-defined tags on an ECS task definition | | Amazon ECS | Amazon ECS Container | User-defined tags on a container running in an ECS task | | Azure | Azure Cost Export - User Resource Tags | User-defined tags on a cloud resource in Azure, found in the **Tags** column in the Azure cost export. Does not include Resource Group tags | | Azure | Azure Cost Export - System Resource Tags | Azure-defined tags on a cloud resource, found in the **AdditionalInfo** column in the Azure cost export | | Google Cloud | Google Billing Export - Project Labels | User-defined labels on a project in Google Cloud, found in the **project.labels** column in the billing export | | Google Cloud | Google Billing Export - System Resource Labels | System-generated labels on a resource in Google Cloud, found in the **system\_labels** column in the billing export | | Google Cloud | Google Billing Export - User Resource Labels | User-defined labels on a cloud resource in Google Cloud, found in the **labels** column in the billing export | | Google Cloud | Google Billing Export - User Resource Tags | User-defined tags on a cloud resource in Google Cloud, found in the **tags** column in the billing export. The `goog-originating-sku-description` tag is added leveraging Google's SKU APIs, to provide more granular SKU details for commitment line items. | | Google Cloud | GKE Pod | User-defined labels found on pods running in Google Kubernetes Engine | | Oracle Cloud | OCI Cost Export - User Resource Tags | User-defined tags on a cloud resource in Oracle Cloud Infrastructure, from the **Tags** column in the OCI FOCUS cost export | | Datadog | Datadog Usage Attribution | User-defined tags for Usage Attribution in Datadog Plan and Usage | | Custom Costs | Cost File Tags | User-defined tags for every provider, found on [cost files](https://docs.datadoghq.com/cloud_cost_management/setup/custom?tab=csv) uploaded to Cloud Cost Management | ## How tags are normalized{% #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 Cloud Cost Recommendations use [standard Datadog metrics normalization rules](https://docs.datadoghq.com/getting_started/tagging/#define-tags). Tag values in recommendations are converted to lowercase and limited to 200 characters. For example, a tag `Team:Engineering-Services` appears as `team:engineering-services` in recommendations, but as `team:Engineering-Services` in cost data. ## Override tag value normalization{% #override-tag-value-normalization %} Turn on **Tag Normalization** in the Tag Pipelines page to normalize all cost tag values to match the Metrics normalization. From the example above, you would see `team:engineering-services` everywhere. Tag normalization applies to user-defined tags from cloud costs. Tags created by Tag Pipelines are not normalized. For Azure, the `consumedservice` out-of-the-box tag is also normalized to lowercase. For all new users, the Tag Normalization toggle is enabled by default, with normalized tag values backfilled for the past 3 months automatically. To backfill normalized tags for a longer period up to 15 months, contact [Datadog support](https://docs.datadoghq.com/cloud_cost_management/allocation/container_cost_allocation/?tab=aws#data-transfer). Tag normalization allows you to: - View, filter and group Cost Recommendations and cost data with the same tag values - Correlate cost and observability data with the same tag values - Automatically combine tag values with different casing (For example: `env:Engineering-Prod` and `env:engineering-Prod` are both normalized to `env:engineering-prod` without having to manually set up Tag Pipelines) {% image source="https://datadog-docs.imgix.net/images/cloud_cost/tag_normalization_toggle.41f43b849e9e0c17afed9d59af7cf488.png?auto=format" alt="Tag Pipelines interface showing optional toggle to normalize all tags on cost data" /%} ## How tags are prioritized{% #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{% #improving-tagging %} 1. **Understand what tags exist** - Use the [Tag Explorer](https://docs.datadoghq.com/cloud_cost_management/tags/tag_explorer) to discover which tags are already available in your cost data. 1. **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. 1. **Split up shared costs** - Use [Custom Allocation Rules](https://docs.datadoghq.com/cloud_cost_management/allocation/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. 1. **Address missing or incorrect tags** - Use [Tag Pipelines](https://docs.datadoghq.com/cloud_cost_management/allocation/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. 1. **Add new tags** - Use [Tag Pipelines](https://docs.datadoghq.com/cloud_cost_management/allocation/tag_pipelines) to automatically create new inferred tags that align with specific business logic, such as a `business-unit` tag based on team structure. {% image source="https://datadog-docs.imgix.net/images/cloud_cost/tag_explorer/aws_1.5b18fa3c369eb5368f143ebf21ebae67.png?auto=format" alt="Tag Explorer interface showing available AWS tags and their usage" /%} ## Further reading{% #further-reading %} - [Learn about Cloud Cost Management](https://docs.datadoghq.com/cloud_cost_management/) - [Tag Pipelines](https://docs.datadoghq.com/cloud_cost_management/allocation/tag_pipelines) - [Tag Explorer](https://docs.datadoghq.com/cloud_cost_management/tags/tag_explorer) - [Getting Started with Tags](https://docs.datadoghq.com/getting_started/tagging/)