Case study · eCommerce
Pay for peak traffic only when it happens
eCommerce platforms are often sized for their busiest day and left that way all year. This is how we cut cloud costs without hurting performance: right-sizing, autoscaling for peaks, committed-use discounts for the steady base and cost visibility that keeps savings from drifting back.
Project overview
Cost optimisation without performance loss.
- Industry
- eCommerce
- Client
- Confidential
- Solution
- Cloud cost optimisation
- Platform
- AWS or Azure
- Timeline
- 6–10 weeks
- Team
- 3 people
- Traffic
- Seasonal peaks
- Governance
- Tagging and budgets
- Reporting
- Cost per team and service
Expected business outcomes
What the client set out to achieve
Savings are measured against the previous months' bills, with performance tracked alongside.
-
20–40%
Lower monthly cloud bill targeted
Right-sizing, autoscaling and commitments combined.
Target range; depends on how the estate was sized before
-
Unchanged
Peak-day performance
Autoscaling tested against peak traffic before changes go live.
Load tested before and after
-
Per team
Cost visibility
Tags and dashboards show who spends what.
Where tagging is enforced
Business challenges
Where the client started
The bill grew with the business, then kept growing.
Sizing
Servers sized for Black Friday, running all year.
Environments
Test environments left running at night and weekends.
Storage
Years of logs and images in the most expensive tier.
Pricing
Everything on pay-as-you-go rates.
Visibility
Nobody knows which team or feature drives the cost.
Proposed solution
What the optimisation covers
Quick wins first, structural changes second, governance throughout.
-
Cost analysis
Spend broken down by service, environment and team.
-
Right-sizing
Instances and databases resized to real usage.
-
Autoscaling
Capacity that grows for peaks and shrinks afterwards.
-
Commitments
Reserved instances or savings plans for the steady base load.
-
Storage tiers
Old data moved to cheaper tiers with lifecycle rules.
-
Governance
Tagging, budgets and alerts so savings don't drift back.
Project requirements
What the client asked us for
The request was a smaller bill. The constraint is peak performance: nothing changes in production until it has been load tested against the busiest day's traffic.
Savings last only with visibility, so every team sees its own spend.
What the project had to deliver
- 01A smaller, predictable bill
- 02Peak performance protected
- 03Costs visible per team
- 04Savings that don't drift back
Key challenges & considerations
What made it harder, and what we did
-
Unknown owners of old resources
What we did
Tagging was enforced, and unowned resources were traced to an owner or removed.
-
Applications that can't scale horizontally
What we did
Applications that could not scale out were right-sized, and the rest were moved to autoscaling.
-
Commitment decisions with growth uncertainty
What we did
Commitments were bought only for the steady base load.
-
Peak dates that limit change windows
What we did
Changes were load-tested and scheduled away from peak dates.
What we watched for
- Load test before resizing production
- Schedule non-production environments
- Buy commitments only for the steady base
- Enforce tagging early
Development process
Measure, save, govern
Changes are scheduled away from peak trading.
-
Analysis
Spend, usage and owners mapped.
-
Quick wins
Idle resources removed and non-production schedules set.
-
Right-sizing and autoscaling
Production resized and autoscaling load tested.
-
Commitments
Reserved capacity bought for the steady base.
-
Governance
Tagging, budgets and dashboards in place.
Technology stack
Cloud-native cost tooling
Native tools first, with dashboards finance can read.
Analysis
- AWS Cost Explorer / Azure Cost Management
- Usage metrics
Scaling
- Autoscaling groups / VM scale sets
- Kubernetes autoscaling
Load testing
- k6 / JMeter
- Peak traffic replays
Governance
- Tag policies
- Budgets and alerts
- Power BI cost dashboards
Quality assurance & deployment
Rehearsed before production, monitored after
Infrastructure changes were proved before they touched production. Every change was made as code, tested in a copy of the environment and rehearsed with a rollback.
- Infrastructure defined as code and reviewed like application code
- Changes tested in a non-production copy first
- Load tests before any resize or cut-over
- Backup restores and failover rehearsed, not assumed
- Security baselines checked automatically
Deployment
- 01CommitmentsReserved capacity bought for the steady base.
- 02GovernanceTagging, budgets and dashboards in place.
- 03Rehearsed cut-oversEach change window was planned, rehearsed and reversible.
- 04Monitoring and handoverDashboards, alerts and runbooks handed to the client's team, and six months of free bug fixing.
Final deliverables
A cloud bill that follows the business
The platform scales up for peaks and back down afterwards, the steady base runs on discounted capacity and old data sits in cheap storage.
Every team sees its own spend, so savings are protected.
- Right-sized resources
- Tested autoscaling
- Committed-use discounts
- Per-team cost visibility
Handed over to the client
- 01Infrastructure as codeIn the client's repository
- 02PipelinesBuild, test and deploy
- 03Monitoring and alertsDashboards and on-call routing
- 04RunbooksFor releases, recovery and failover
- 05Six-month warrantyFree bug fixing after delivery
Success metrics
How success is measured
| Measure | Outcome | How it is measured |
|---|---|---|
| Lower monthly cloud bill targeted | 20–40% | Target range; depends on how the estate was sized before |
| Peak-day performance | Unchanged | Load tested before and after |
| Cost visibility | Per team | Where tagging is enforced |
Verified reviews
What clients say about our custom software work
Verified reviews from our clients on the same kind of work, published on Clutch.
-
Custom Software Development
Web-Based Coupon App Development
The work produced by the team was of exceptional quality.
Jon Thies
CTO, Model Rocket Feb 2022 – Mar 2023
-
Custom Software Development
Mobile & Web Platform Development & Design
Silver Scintilla was responsive, respectful, and enjoyable to work with, quickly addressing any questions or requests.
DeShawn Brown
CEO & Founder, Lithios Jun 2021 – Nov 2022
-
Custom Software Development
Web-Based HRM System Development
Silver Scintilla provided the best solution with scalable functionality, personalized features, and robust security.
Suheb Khan
Director, Collaborative Insight Technologies Jul 2020 – Jun 2022
Free discovery call
Have an idea? Let's turn it into AI-powered software.
Book a free discovery call with our experts. Share your idea and we will help you shape the scope, timeline and budget, under NDA.
- Free consultation
- NDA before we talk
- Transparent estimate