Rightsizing recommendations grounded in real usage
The agent compares what you pay for compute with what services actually use in CloudWatch or Datadog, then recommends instance sizes to the owning team, with the deploy history attached.
Runs in
Claude · Slack
First optimization recommendation
The next day
ROI
Pays for itself with the first recommendation you act on
Which EC2 services ran below 30% average CPU over the last 30 days according to CloudWatch? Group them by owner and include recent deploys.
api-workers (Checkout) runs on 12× m6i.4xlarge at 22% average CPU, p95 41% (Datadog, 30 days). Moving to m6i.2xlarge keeps p95 under 85%. Estimated impact: about €3,100/month. No deploy in the last 14 days changed its load profile.
Next step: I can set up an alert on api-workers, so you hear about it if its CPU stays this low. Should I create it?
Yes, set up that alert.
Alert created for api-workers. It is checked after each data refresh and posts to your Slack channel with a link to Explorer.
The owning team decides whether to change the instance size.
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
A service runs well below its capacity for 30 days.
Costory investigates
01
Join cost and usage
Billing lines with compute metadata sit next to CloudWatch or Datadog utilization.
02
Add context
Owners from Virtual Dimensions and deploys from GitHub or CI/CD events explain recent changes.
03
Claude proposes
Candidate size changes per service, with peak usage, owner and the estimated monthly impact.
04
The team decides
Recommendations go to the owner's channel. Costory never changes your infrastructure.
Costory suggests
Next step: I can set up an alert on api-workers, so you hear about it if its CPU stays this low. Should I create it?
Where you see it
Claude
You decide
The owning team decides whether to change the instance size.
Why it matters
Generic rightsizing lists ignore context: the service that peaks on Mondays, the migration in progress, the team that owns the box. Engineers ignore recommendations they can't trust.
Who it’s for
Platform engineers, SREs, engineering managers
Where it runs
Claude · Slack
What you get
What lands in front of your team
Evidence attached
Each recommendation shows usage, cost and recent deploys.
Routed to the owner
The team that runs the service sees it first.
Humans decide
No automated changes, no write access to your cloud.
Sources and assistants
What it connects
Data sources
Example prompts
Or drive it yourself
Find
Which EC2 services ran below 30% average CPU over the last 30 days according to CloudWatch? Group them by owner and include recent deploys.
Recommend
For the api-workers service, propose a rightsizing option that still covers the p95 load from Datadog, with the monthly impact.
Share
Post the top 5 rightsizing candidates for the Checkout team to #team-checkout with the evidence.
FAQ
Questions about this playbook
Does Costory apply the change?
No. Claude recommends and the owning team decides. Costory has no write access to your cloud accounts and does not apply automated rightsizing.
Where does utilization data come from?
From the usage metrics you connect, such as AWS CloudWatch, Datadog or GCP Cloud Monitoring, queried next to cost.
Does Costory guarantee savings?
No. The agent investigates and recommends with the evidence attached; your team decides what to change. Impact figures are estimates based on your own data.
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
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
Costory notices
Monthly review of spend on older instance families.
Find candidates for newer instance generations and Graviton
Use the compute metadata in your billing data to list older instance families by owner and estimate the impact of upgrading.
See the playbook
Costory notices
An optimization is logged as shipped.
Measure the impact of every optimization you ship
Log each optimization as an event and compare cost before and after, so the team can show what each change saved.
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