Rightsizing
Rightsizing reduces cost by matching resource size and configuration to actual usage. This guide explains how to use Cloudaware FinOps to identify rightsizing opportunities, validate recommendations, and reduce spend by aligning resource capacity with actual workload demand.
What to Rightsize
Common candidates include:
Compute instances (VMs, instances, nodes) with consistently low CPU, memory, or I/O utilization.
Managed databases and analytics services sized well above observed load.
Storage classes or volumes where performance tiers exceed requirements.
Focus first on high‑spend resources and critical applications where savings will be meaningful and safe.
Inputs and Signals
Rightsizing recommendations can draw on:
Cloud provider metrics (for example, CPU, memory, network, IOPS where available).
Breeze Agent and CMDB attributes (OS‑level utilization, roles, attached storage, dependencies).
Cost metrics such as effective hourly rate and commitment coverage.
Use tags and business mapping to view candidates by application, BU, environment, or customer so owners can prioritize and act.
Workflow
A typical rightsizing loop:
Discover candidates.
Use dashboards and reports to list low‑utilization resources by scope (BU, app, environment).
Filter out short‑lived or intentionally over‑provisioned resources.
Validate with owners.
Confirm workload characteristics with application teams (for example, bursts, upcoming events, latency requirements).
Adjust thresholds or filters if many false positives appear.
Plan changes.
Choose target sizes or families and maintenance windows.
For stateful services, plan backups and rollback options.
Execute and monitor.
Apply changes via your preferred automation (Cloudaware automation, cloud native tools, or IaC pipelines).
Monitor performance and cost after the change; be prepared to roll back if needed.
Track savings.
Capture realized savings in dashboards and reports to demonstrate impact over time.
Integrate rightsizing work with your broader optimization playbooks and change management processes so recommendations turn into consistent action.
Rightsizing in Cloudaware
In Cloudaware, rightsizing signals combine provider metrics, CMDB context, and (where available) Breeze Agent telemetry to identify over‑ and under‑provisioned resources.
Cloudaware also uses Compliance Engine policies to generate policy-driven recommendations for resources such as AWS EC2 instances. For example, it can recommend downsizing overprovisioned EC2 instances to help reduce cloud costs. These recommendations are available within the CMDB.