Documentation and Best Practices

Learn how to use Cloudability and get the most out of our cloud cost management tool.

Follow

Making the most of popular dimensions and metrics within Cloudability

There are a vast number of dimensions and metrics available in Cloudability reporting. We've listed some of the most popular and useful below. For a full list, check out our Glossary of Dimensions and Metrics.

 

Service Name (vs Product Name)

Cloudability created Service Name to standardize mixed naming conventions for AWS and Azure's Product Name field:

  • AWS EBS becomes its own top-level product, as opposed to being considered EC2 in the DBR file.
  • Unified data that was mixed across multiple products:
    • AWS tracks Cloudwatch partially as a product and as EC2; this separates Cloudwatch into its own service
    • AWS tracks Glacier partially as a product and in S3; this consolidates all Glacier and S3
    • AWS has an import/export service, as well as a Snowball service; Service Name consolidates these
  • In some cases, AWS and Azure data shows multiple products for the same services:
    • AWS has multiple names for Support, this consolidates them all into a single "AWS Support" line item
    • AWS has many administrative charges listed as top-level "products" like: Amazon Partner Network, re:Invent Tickets, etc. These are consolidated into a single "AWS Administrative" item
    • Azure has Compute, Microsoft.Compute, and Virtual Machines; these are all the same and consolidated under Azure Compute.
  • AWS billing files have inconsistent naming schemes for products, sometimes Amazon vs AWS; acronyms vs spelled out names. Service Name follows a single pattern of a prefix of "AWS" followed by the name that is seen in the AWS Console menu: EC2, S3, RDS, Kinesis, Lambda, etc. Similarly, Azure is the prefix for all Microsoft Azure services.

Also, when Cloudability migrated our data source to the Detailed Billing Report (DBR), the Product Code dimension was deprecated. Any saved reports should should use Service Name. 

Usage Family (vs Usage Type)

Usage Type is a very granular dimension that exposes specific operational details of each line item. This dimension can provide a great amount of utility when combined with 'contains' and 'does not contain' filter parameters.

Typical values to see returned are 'BoxUsage:c3.xlarge' to signify on-demand usage, or 'HeavyUsage:m2.4xlarge' to signify Heavy/Reserved usage, or 'EBS:VolumeP-IOPS.piops' to signify Provisioned IOPS usage.

A usage type of 'Unknown' is representative of things like the sign up charge from an RI purchase.

Usage Family, in turn, is a standardized dimension created by Cloudability to consolidate similar values of Usage Type into a single high-level family, e.g. Instance Usage, Data Transfer, Storage, or API Request. It is a great place to start breaking down your spend beyond Accounts and Services.

 

Item Description

Item Description is an even more granular dimension that can expose still further operational detail of each line item. This dimension can also provide a great amount of utility when combined with ‘contains’ and ‘does not contain’ filter parameters.

Typical value to see returned are ‘$0.105 per On Demand Linux c3.large Instance Hour’ to signify on-demand usage, or ‘$0.070 per GB - next 100 TB / month data transfer out’ to signify data transfer, or ‘USD 0.5589 hourly fee per Linux/UNIX, i2.4xlarge instance’ to signify RI usage hours, or ‘$0.0290 per GB - next 450 TB / month of storage used’ to signify S3 costs, or ‘$0.850 per Redshift Dense Storage Extra Large (DW1.XL) Compute Node-hour (or partial hour)’ to signify Redshift costs.

An Item Description of ‘Sign up charge for subscription: ####, planId: ###’ (where ### are actual numbers) is representative of sign up charges from an RI purchase. This allows you to create a filter for ‘item description_does not contain_sign up’ to see your compute costs without the upfront fee included.

 

Transaction Type

This dimension was created by Cloudability to discretely break out Recurring Charges like AWS Support and monthly RI fees, One-Time Charges like RI sign up payments, Credits, Taxes, and Usage costs. It can be tremendously helpful in determining your RI efficiency and waste, and also helps to show how Blended and Unblended cost can vary by line item, even when they reach the same month-end total. 

 

Compute Usage Type

This Dimension very clearly breaks out what instances are running on-demand, spot, reserved, an RDS node, etc. Running a Cost Analytics report with ‘Compute Usage Type’ and ‘Item Description’ as the only Dimensions will surface a complete detailed view.

In order to see only your compute costs, add a filter for ‘compute usage type_not equals_other’ as the ‘other’ usage type represents costs such as data transfer, elastic IP mapping, alarm months, read /write hours, GB storage, LoadBalancer GB processed, shard-hours, PIOPS, magnetic storage, and others.

In order to see your Redshift costs, add a filter for ‘compute usage type_equals_node’.

In order to see your ElastiCache costs, add a filter for ‘compute usage type_equals_node usage’.

In order to see your RDS costs, add a filter for ‘compute usage type_equals_instance usage’.

In order to see your Heavy RI costs, add a filter for ‘compute usage type_equals_heavy usage’.

In order to see your On-Demand costs, add a filter for ‘compute usage type_equals_box usage’.

  

On-Demand Hours, Reserved Hours, Reserved Utilization Rate

These Metrics are invaluable in determining exactly which accounts, services, tags, or resources leveraged your Reserved Instances. The Detailed Billing Report is populated with hourly resolution across all of these parameters, allowing our analytics platform to group by any Dimension and show how it took advantage or RI hours over any reporting period chosen.

Adding a filter for ‘reserved utilization rate_greater than_X%’ (where ‘X%’ is a number without the percent symbol) is a great way to see which areas of your infrastructure could benefit from additional RI coverage.

 

Instance Type

This Dimension combines the Instance Family and Instance Size Dimensions and presents the full identifier of the instance, eg c3.large.

In Cost Analytics, in order to see only your compute costs, add a filter for ‘instance type_not equals_none’ as the ‘none’ instance type represents data transfer, elastic IP mapping, alarm months, read /write hours, GB storage, LoadBalancer GB processed, shard-hours, PIOPS, magnetic storage, GET /PUT /POST /COPY /LIST requests, and others.

In Usage Analytics this Dimension will return only your EC2 instances.

 

Running Instances and Average Running Instances per Hour

The Running Instances Metric is a count of all unique instance ids that were running at one point during the reporting window. This is not necessarily a count of the instances that are live or running at the time the report is run - rather the sum of unique instance ids that were running.

Average Running Instances per Hour is an average of the number of instances that were live each hour of the reporting window.

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request

0 Comments

Please sign in to leave a comment.
Powered by Zendesk