Independent studio · DevOps · Cloud
The engineer,
not the bench.
When you hire Drozan you get that engineer. There is no account manager, no junior doing the real work, and no deck that arrives three weeks after the kickoff. Work lands as pull requests in your repositories from the first week.
- In production
8 years
Pipelines, clusters, networks, bills.
- Where
India / worldwide
IST, overlapping European afternoons and US mornings.
- Clouds
AWS · GCP · Azure
Production, not labs.
Consulting
DevOps and cloud engineering for your team
Architecture reviews, Kubernetes platforms, CI/CD rebuilds, observability, cost and migrations. Senior-level ownership without the cost and delay of a full-time hire.
Studio
Mobile apps published under our own name
Drozan is the umbrella for the apps we build ourselves. The first is in progress. Running our own releases, on-call and cloud bill means the advice we give has already been paid for once, by us.
Where we come in
Seven areas. Most of what goes wrong in a modern engineering organisation.
- 01
Cloud & Infrastructure Architecture
Landing zones, account structure, networking and IaC foundations designed so the next fifty engineers inherit something sane instead of something archaeological.
- 02
Kubernetes & Container Platforms
Production-grade clusters with the boring parts automated: ingress, autoscaling, GitOps delivery and a clear blast radius for every workload.
- 03
CI/CD & Developer Experience
Pipelines your team trusts. Fast feedback, reproducible builds, and releases that are a non-event rather than a calendar invite.
- 04
Observability & SRE
Know what broke before your customers tell you. Metrics, logs and traces wired into SLOs and an on-call rotation people can actually live with.
- 05
Cloud Cost Engineering
Most cloud bills are 30-50% waste hiding behind unlabelled resources. We find it, cut it, and put guardrails in so it does not creep back.
- 06
Security & Compliance Engineering
Least-privilege IAM, real secret management and supply-chain controls built into the pipeline, so compliance becomes a byproduct instead of a project.
- 07
Migration & Modernization
Data centre to cloud, VM to container, monolith to services — sequenced so the business keeps shipping while the platform changes underneath it.
Engagements
Three ways to work together.
Most clients start with a review and move into a retainer once the roadmap is clear. Pricing depends on scope — you get a concrete proposal, not a rate card.
Engagement
Architecture Review
1-2 weeks, fixed scope
You inherited a platform and need an honest read on it.
- Infrastructure, pipeline and security audit
- Prioritized findings with effort and impact
- Cost reduction opportunities, quantified
- Written remediation roadmap and walkthrough
Most common
Fractional DevOps
Monthly retainer, agreed hours
You need senior platform ownership without a full-time hire.
- Ongoing platform ownership and on-call escalation
- Roadmap, standards and code review for infra changes
- Mentoring for your existing engineers
- Cost and reliability reporting each month
Engagement
Project Delivery
Fixed scope, milestone-based
A defined outcome: a migration, a platform build, a rebuild.
- Discovery, design and a signed-off delivery plan
- Implementation with your team in the loop
- Documentation, runbooks and handover
- Post-delivery support window
Method
Four steps, no mystery.
Every engagement runs the same way, whether it is a two-week review or a six-month platform build.
- 01
Discovery
A working session on your architecture, pipelines, incidents and bills. No slide deck required — access and honesty are enough.
- 02
Assessment
A written read of what is solid, what is fragile and what is costing you money, with everything ranked by impact against effort.
- 03
Build
Changes land as reviewed infrastructure code in your repositories, your cloud, your pipelines. Nothing is built behind a curtain.
- 04
Handover
Runbooks, architecture decision records and working sessions so your team owns the platform when the engagement ends.
Writing
Notes from production.
What actually breaks, what it costs, and what we would do differently.
Field note · Cloud cost
The $438 cluster nobody approved
EKS extended support costs six times standard support, starts automatically, and nobody has to approve it. How to check whether you are paying it right now.
20 September 2026
5 min readField note · Cloud cost
Paying twice to pull the same image
NAT Gateway charges $0.045 per GB. Every image pull from a private subnet goes through it. What that costs at scale, and the endpoints that stop it.
20 September 2026
5 min readField note · Cloud cost
Where your EKS bill actually goes
The compute line item is the one everybody looks at. It is rarely the one that grew. Five charges that hide inside a Kubernetes cluster, and how to find them.
20 September 2026
6 min read
Start here
Tell us what is breaking.
A slow pipeline, a cloud bill nobody can explain, a cluster that only one person understands. Send the details and you will get a straight answer on whether we can help.