AWS Well-Architected Sustainability Pillar
The Sustainability Pillar gives cloud teams a repeatable method to cut the carbon footprint and cost of their workloads by maximizing utilization and minimizing wasted capacity. It turns environmental responsibility into concrete architecture and operations decisions.
Best Practice: AWS Well-Architected Sustainability Pillar
The Sustainability Pillar is the sixth pillar of the AWS Well-Architected Framework, added in 2021. It gives teams a structured way to understand and reduce the environmental impact of running cloud workloads. The core idea is straightforward: every unit of compute, storage, and network you provision but do not use carries an environmental and financial cost. By maximizing utilization and minimizing the total resources required to do useful work, you lower both your carbon footprint and your bill. This matters because cloud spend and energy use scale with provisioned capacity, not delivered value, so waste is common and often invisible.
Step-by-Step Implementation Guidance
- Establish a baseline. Measure current utilization across compute, storage, and network using cloud monitoring data so you know where waste lives.
- Set sustainability goals tied to business metrics, such as resource use per transaction or per active user, rather than absolute totals.
- Right-size workloads. Remove idle and over-provisioned instances, volumes, and clusters; use autoscaling to match capacity to demand.
- Choose efficient regions and services. Prefer regions with lower-carbon energy and managed or serverless services that share infrastructure efficiently.
- Optimize data. Apply lifecycle policies, tiered storage, compression, and deduplication; delete data nobody reads.
- Adopt efficient hardware. Use modern instance families (for example Graviton) and accelerators sized to the job.
- Review continuously. Treat sustainability as an ongoing operational practice with regular checks, not a one-time audit.
Common Mistakes Teams Make When Ignoring This Practice
- Provisioning for peak load permanently instead of scaling elastically.
- Leaving development, test, and orphaned resources running outside business hours.
- Storing all data on hot, high-performance tiers regardless of access patterns.
- Treating sustainability as a reporting exercise disconnected from architecture decisions.
- Ignoring the carbon profile of the regions they deploy to.
Tools and Techniques That Support This Practice
- AWS Compute Optimizer and Trusted Advisor for right-sizing recommendations.
- Amazon CloudWatch for utilization metrics.
- AWS Customer Carbon Footprint Tool for emissions reporting.
- S3 Storage Lifecycle policies and Intelligent-Tiering for data efficiency.
- AWS Graviton-based instances for energy-efficient compute.
How This Practice Applies to Different Migration Types
- Cloud Migration: Right-size workloads during the move instead of lifting-and-shifting oversized on-premise capacity.
- Database Migration: Choose managed engines and storage tiers matched to real query and retention patterns.
- SaaS Migration: Favor multi-tenant SaaS that amortizes infrastructure across customers over self-hosted equivalents.
- Codebase Migration: Refactor toward serverless and event-driven patterns that consume resources only when work runs.
Checklist
- Baseline utilization across compute, storage, and network.
- Set sustainability targets tied to a business unit of work.
- Right-size and enable autoscaling for all production workloads.
- Apply storage lifecycle and tiering policies.
- Adopt energy-efficient instance families where supported.
- Schedule non-production environments to shut down when idle.
- Review sustainability metrics on a recurring cadence.