Dynamic cost reallocation for EKS using Datadog usage
Split shared EKS cluster costs across teams in proportion to what each team's workloads actually use, measured in Datadog.
Runs in
Claude · Slack
First optimization recommendation
The next day
ROI
Pays for itself with the first recommendation you act on
Playbook
What the agent does
From what Costory notices to what you decide. Costory suggests the next step in conversation, and you decide whether to go ahead.
Costory notices
Daily, as Datadog usage and the EKS bill refresh.
Context used
Remembered
Costory investigates
01
Connect AWS and Datadog
CUR 2.0 brings the EKS cost; Datadog is connected as an external metric source next to billing.
02
Pick the usage signal
Choose the Datadog metric that best reflects consumption, for example CPU or requests per service.
03
Map labels to teams
Identity, regex or manual mapping turns metric labels like team-payments-prod into team names.
04
Reallocate and publish
The shared cost is split in proportion to usage, and the split follows usage as it moves. Doing this in a DIY setup means building and maintaining a metrics pipeline.
Where you see it
Claude
You decide
Platform approves the split method once; teams see their share every day.
Why it matters
Shared clusters make one big line on the AWS bill. Splitting by node count or headcount feels unfair, and teams stop trusting the numbers. Usage-based splits fix that, but keeping them current by hand is painful.
Who it’s for
Platform teams, Kubernetes owners, FinOps leads
Where it runs
Claude · Slack
What you get
What lands in front of your team
A fair split
Each team pays for the share of the cluster its workloads used.
Always current
The split follows usage instead of a yearly spreadsheet.
Transparent method
The usage driver and mapping rules are visible to every team.
Sources and assistants
What it connects
Example prompts
Or drive it yourself
Design
Our prod-eu EKS cluster is shared by five teams. Propose a usage-based split using Datadog CPU usage per service and show what each team would have paid last month.
Check
Compare the current headcount-based split with a Datadog usage-based split for the shared cluster. Who pays more, who pays less?
Share
Post the new usage-based split for prod-eu to #platform with a short explanation of the method.
FAQ
Questions about this playbook
Which usage metrics can drive the split?
Supported metric sources include Datadog, Datadog CI Visibility, BigQuery, Tsuga, AWS CloudWatch, S3 metrics, Amplitude, Google Sheets, Snowflake and custom pushed metrics.
Does this replace AWS Split Cost Allocation Data?
No, it complements it. Split Cost Allocation works at pod level inside EKS; usage-based reallocation helps with shared costs that pod labels don't explain.
What if our tagging is incomplete?
Costory's Virtual Dimensions allocate with rules on accounts, projects, names and partial tags, so you don't need perfect tagging to start. Shared and untagged cost is reallocated dynamically using the metrics you already have (Datadog, Tsuga, BigQuery, Snowflake and more), and the allocation stays consistent as usage moves.
Does Costory change anything in our cloud accounts?
No. Costory reads billing exports and the metrics you connect, and has no write access to your cloud. Actions it takes create Costory objects only: alerts, reports, dashboards and events. Your team stays in control of every change.
Step-by-step guide in the Costory docs
Related playbooks
Run these alongside it
Costory notices
Every day, for every cluster and namespace.
Kubernetes cost allocation by namespace, team and workload
Break EKS, GKE and AKS costs down by cluster, namespace and workload, and roll them up to teams.
See the playbook
Costory notices
A shared service's cost grows, or a new team starts using it.
Split shared infrastructure costs by actual usage
Allocate shared databases, networking and platform services to the teams that use them, based on real usage metrics.
See the playbook
Costory notices
Idle capacity in a node pool stays above your threshold.
Find over-requested Kubernetes workloads and right-size requests
Measure the gap between requested and available CPU and memory, convert it to cost, and get request changes per deployment.
See the playbook
Put this playbook on call
Included in every plan, from €250/month. 30-day free trial: connect in under 30 minutes, get your first optimization recommendation the next day.
First optimization recommendation
The next day
ROI
Pays for itself with the first recommendation you act on
Setup
About 30 minutes
Data residency
EU
Access
Read-only on your cloud