DevOps
How we cut a fintech's cloud bill by 42% and reduced deploy time from 45 minutes to 8 — in 6 weeks.
- DevOps team
- 6 weeks
- 4 engineers
- Live in production
Cloud cost reduction
0
%
Deploy time (was 45min)
0
Min
Failed releases post-launch
0
Uptime achieved
0
%
The challenge
A fast-growing fintech paying $380K/year in cloud costs — for infrastructure delivering only 60% of what it should.
The engineering team was shipping 2–3 releases per week, each requiring a 45-minute manual deployment process with four engineers on a call. One bad release every month caused customer-facing downtime averaging 90 minutes.
Meanwhile their AWS bill was growing 15% month-on-month, but the CTO had no visibility into what was driving it or where to cut without risking stability.
What we found in the audit :
- $140K/year in over-provisioned EC2 instances running at 12% utilisation.
- No CI/CD — all deployments manual, error-prone, and undocumented.
- Zero monitoring or alerting — outages discovered by customers first.
- No rollback strategy — every bad deploy meant a 2-hour all-hands incident.
Project details :
Industry
FinTech
Team size
120 engineers (client)
Duration
6 weeks
Our team
4 DevOps engineers
Status
Live
Technology stack
- GitHub Actions
- AWS
- Terraform
- ArgoCD
- Kubernetes
- Datadog
Our solution
Full infrastructure rebuild — audit, IaC, CI/CD, monitoring — in 6 weeks flat.
We started with a brutal audit, then redesigned the infrastructure from the ground up — rightsizing everything, codifying it all in Terraform, and replacing manual deploys with a fully automated GitHub Actions → ArgoCD pipeline.
What we built :
Full infrastructure audit — week 1
Catalogued every AWS resource, mapped costs, identified $160K/year in savings opportunities across compute, storage, and data transfer.
→ Audit report + savings roadmap
Infrastructure as code — weeks 2–3
Rewrote entire infrastructure in Terraform. Rightsized 34 EC2 instances. Migrated to reserved instances for predictable workloads.
→ Full Terraform codebase + cost model
CI/CD pipeline — weeks 3–4
Built GitHub Actions → ArgoCD pipeline with automated testing, canary deploys, and one-click rollback. Deploy time dropped from 45 min to 8 min.
→ Live CI/CD pipeline + runbook
Observability stack — weeks 5–6
Deployed Datadog across all services — custom dashboards, SLO alerting, anomaly detection, and on-call runbooks for every alert.
→ Monitoring dashboards + alert playbooks
The results
14 weeks. Zero disruption. Transformational outcomes.
42%
Lower cloud bill
$160K
Annual savings
8 min
Deploy time
0
Failed releases
Why clients trust us
We were spending $380K a year on AWS and had no idea what we were getting for it. In 6 weeks Benthon cut the bill by 42%, gave us a CI/CD pipeline that actually works, and set up monitoring so we find problems before our customers do.
John Doe
Product Manager, Indyfin
Explore more case studies
Transforming Patient Care with Advanced Telemedicine Solutions
Created via a full spectrum of IT services, extensive technical expertise, and well-adjusted solution design
- EdTech
Streamlining Financial Transactions with Blockchain Technology
Created via a full spectrum of IT services, extensive technical expertise, and well-adjusted solution design
- EdTech
Streamlining Financial Transactions with Blockchain Technology
Created via a full spectrum of IT services, extensive technical expertise, and well-adjusted solution design
- Compliances
Boosting Online Sales with Personalized Shopping Experiences
Created via a full spectrum of IT services, extensive technical expertise, and well-adjusted solution design
- EdTech