Home › Guides › IT Support to Cloud Engineer
Career & Salary · 2026IT Support to Cloud Engineer in India 2026: The Real Timeline, Skills Gap and What It Pays
Moving from IT support to cloud engineer in India takes 10 to 14 months of part-time study alongside your current job. You close four gaps: Linux depth, networking, one cloud platform certified at associate level, and infrastructure as code. First cloud-facing roles are advertised around 4 to 9 LPA, so expect a flat move before the jump.
- Your ticket history is an asset, but only if you translate it. On-call instinct, SLA discipline and log reading transfer directly. The words on your resume do not.
- There are four gaps, not ten. Linux without a runbook, networking below the desktop layer, one cloud platform at associate depth, and infrastructure as code. Everything else is optional for the first role.
- The first cloud role usually pays sideways. Entry cloud and cloud support roles are commonly advertised in the 4 to 9 LPA band, which for an L2 engineer on 6 LPA is a lateral move, not a raise.
- One associate certification is enough to get read. SAA-C03 costs $150, roughly ₹15,135 from India once 18% GST is added, and the pass mark is 720 out of 1000.
- Demand is real but junior-light. Naukri listed 14,354 cloud engineer vacancies across India when checked on 12 September 2026, and most of them ask for two years of cloud experience you do not have yet.
- The money arrives at the second move, not the first. Mid-level cloud engineers with two to four years in the role are typically advertised in the 8 to 16 LPA band.
It is 2:40am, you are three hours into a rotational shift, and the ticket says an EC2 instance is unreachable. You restart the service, close the ticket, write two lines in the handover doc. Somebody else built that instance. Somebody else chose its security group, and somebody else will get the call when it has to scale. That gap, between operating the thing and owning it, is the entire distance between where you sit now and a cloud engineer role. It is a real distance. It is also shorter than most people on LinkedIn make it sound, and this guide is about how long it actually takes and what it costs you on the way.
Is the Switch From IT Support to Cloud Engineer Realistic for You?
Most switch guides skip this part because the honest answer loses readers. Here is the filter I would apply before spending a rupee.
Three signals that say go
You already have shell or RDP access to something in production, even if the runbook tells you exactly what to type. You have written something that runs on its own, a PowerShell script, a scheduled job, even a long Excel macro. And you can genuinely protect eight to ten hours a week for a year, which usually means giving up one full weekend day, not finding magic hours on weeknights after a shift.
Two signals that say wait
If your role is pure ticket routing with no access to any system, you are not switching to cloud engineering in a year. You are switching to systems work first, which is a different and shorter project. And if the real problem is that you are underpaid for support work rather than bored by it, a cloud transition is an expensive way to fix a pay band that a straight job change could fix in eight weeks.
Here is the opinion I will commit to: if you can only fix one thing before you start, fix Linux, not certifications. A candidate who can explain why a process is in D state and fix a full /var partition under pressure gets hired over a candidate with a certification and no hands. The certification opens the screening call. Linux survives the technical round.
Three routes out of the service desk
The internal move is the fastest to start and the slowest to pay. The junior engineer route is the opposite.
Route timings are typical ranges reported by career switchers, not guarantees. Advertised pay band checked 12 September 2026.
The Skills Gap: What Support Work Already Gave You
Take the example we will keep coming back to: a 27 year old L2 service desk engineer at a managed services provider in Pune, four years in, 5.8 LPA, on a rotational night shift. On paper that person has zero cloud experience. In practice they have about 40% of what a junior cloud engineer needs, and it is the 40% that is hardest to teach.
| What service desk already gave you | What a cloud engineer job description asks for | Gap size and rough time to close |
|---|---|---|
| Incident triage, escalation judgement, SLA discipline | On-call ownership, blameless postmortems, error budgets | Small. Two to four weeks of reading plus one runbook you wrote yourself |
| Reading logs until the ticket closes | Reading logs to change a system: CloudWatch, structured logs, metrics, traces | Medium. Six to eight weeks |
| Excel or ServiceNow reporting | Python or Bash that other people run unattended | Medium. Eight weeks |
| Linux or Windows commands copied from a runbook | Linux without the runbook: systemd, permissions, disks, journalctl, process states | Large. Ten to twelve weeks |
| AD, DNS and DHCP from the desktop side | TCP/IP, DNS, TLS, routing, VPC subnets, NAT gateways, security groups | Large. Eight to ten weeks |
| Usually nothing | One cloud platform at associate depth: the SAA-C03 or AZ-305 path | Large. Ten to fourteen weeks |
| Usually nothing | Terraform, Git, one CI pipeline, Docker basics | Large. Ten to twelve weeks |
Those last four rows are the job. Notice how closely they match what the hyperscalers themselves screen for: the AWS Cloud Support Associate graduate posting circulating for the 2026 India intake asks for strong Linux fundamentals, working knowledge of TCP/IP, DNS, HTTP and TLS, and scripting in Python or Bash. No Kubernetes, no Terraform, no machine learning. That posting is a fair statement of the true entry bar, and it is lower than the internet suggests.
Indian job descriptions for the role one rung up look different. Cloud engineer listings in 2026 cluster around Terraform, Kubernetes, Git, observability and infrastructure as code, which is why the plan below spends its back half there rather than on more certifications. 360DT's live AWS Solutions Architect and DevOps Engineer course works through that exact stack on weekends, building the VPC, the Terraform modules and the pipeline in class rather than describing them.
Also read: AZ-305 vs SAA-C03 in 2026: Which Cloud Architect Certification Is Better for Indian Careers?
IT Support vs Cloud Engineer Pay in India in 2026
This is the section people scroll to, so read the caveat first. Indian salary data is messy. Glassdoor's India page for L1 Technical Support Engineer shows an average near ₹7.5 LPA with a 25th to 75th percentile spread of roughly ₹3.4 L to ₹13 L, which tells you more about how wide and how self-reported that band is than about what any one service desk pays. Treat every figure below as an advertised or reported range, not as an offer.
Advertised and reported pay along the path
The first cloud role is not where the money is. The second one is.
Collection basis: advertised ranges and reported averages seen on Naukri, Glassdoor India and 2026 salary round-ups published by Indian hiring platforms including Instahyre and CloudThat, checked 12 September 2026. Bars are drawn from the midpoint of each range against a 40 LPA scale. These are market observations, not guarantees of individual pay.
The shape matters more than any single number. Your first cloud-facing role probably pays what you earn now, sometimes slightly less. The step that changes your life financially is the one after it, once you have two years of cloud on your record. The top of the band is set by three things and none of them is a certification: whether you are at a GCC or a product company rather than a services vendor, whether you own infrastructure as code rather than consoles, and whether you can talk about cost. An engineer who can walk into a review and say which three line items are 60% of the AWS bill is worth measurably more than one who cannot.
For context on how the money runs once you are established, see our breakdown of DevOps engineer pay in India by experience and company type, which is the band most cloud engineers converge on by year four.
A 12-Month Plan to Go From IT Support to Cloud Engineer
Eight to ten hours a week, every week, for a year. That is the whole input. The deliverable column is the part that matters, because a month with no artefact at the end of it did not happen.
| Months | Focus | Concrete deliverable | Hours a week |
|---|---|---|---|
| 1 to 2 | Linux and networking without a runbook | A three VM lab on your own laptop: nginx behind a reverse proxy, a TLS certificate you installed by hand, and a written note on what broke | 8 |
| 3 to 4 | One cloud platform, wide before deep | A VPC built by hand with public and private subnets, a NAT gateway, a bastion host, and a one page cost estimate for running it | 8 to 10 |
| 5 to 6 | Certification sprint | SAA-C03 booked and sat. Pass mark is 720 out of 1000, so do not book until practice scores sit near 800 | 10 |
| 7 to 8 | Infrastructure as code and Git | The same VPC rebuilt in Terraform, in a public repo, destroyed and recreated from zero with one command | 8 |
| 9 to 10 | Containers and CI | A small API in Docker, deployed by a GitHub Actions pipeline, with a rollback you have actually tested by breaking it | 8 |
| 11 to 12 | Evidence and applications | A README-led portfolio, a rewritten resume, ten applications a week, and one direct conversation with the cloud team in your own company | 6 |
Two deliberate choices in that plan. The certification sits in the middle, not at the start, because a certification earned before you have touched anything reads as theory in an interview and you will be caught in the first ten minutes. And Kubernetes is absent, because it is not on the entry bar and it will eat four months you do not have. Learn it in your first cloud job, on someone else's cluster, at their cost.
If your workplace runs Microsoft rather than AWS, swap the platform and keep the structure. The Azure Solutions Architect and DevOps Engineer track covers AZ-305 and AZ-400 in the same order, and the AZ-305 exam is priced for India at roughly ₹4,800 plus tax at Pearson VUE checkout, which is a materially lower barrier than the $150 AWS fee. If you want to compare the full set of exams before committing, the certifications overview lays them out side by side, and the free webinars are a cheaper way to find out whether the subject actually holds your attention for a year.
What You Actually Earn During the Transition
Nothing, for twelve months. That is the first honest answer. Your study year is unpaid, it costs you roughly ₹15,000 to ₹16,000 in exam fees plus whatever training you buy, and it costs you most of a weekend for a year.
The second answer is about the offer itself. If you move internally, expect a title change and a correction in the 0 to 10% range, because internal moves are priced against your current CTC. The Michael Page India Salary Guide 2026 puts general increments in the 8 to 12% range, with promotions closer to 20 to 30% and emerging-skill roles higher, so an internal cloud move at 10% is normal rather than insulting. If you move externally into a first cloud role, you are being priced as a candidate with zero cloud years, which is why the 4 to 9 LPA band bites.
Here is the line I would use when the recruiter asks for your expectation on that first external cloud move, and you know the band is tight:
"I am on 5.8 fixed today in an L2 support role. I know I am entering cloud engineering without production cloud years, so I am not asking you to pay for experience I do not have. What I want is a flat move, not a cut: hold me at 5.8 to 6.2 fixed, and I would rather have the on-call rotation and the Terraform ownership than another 50,000 on the base. Can we look at a six month review written into the offer?"
That works because it is true, it prices you honestly, and a written six month review is far easier for a hiring manager to approve than a number outside the band. Take the flat move when the role gives you infrastructure to own. Refuse the pay cut when the role is support with the word cloud in the title, which is most of what you will be offered in months eleven and twelve.
A Composite Transition, Start to Finish
This is an illustrative composite built from typical market paths, not a real named person and not a 360DT student outcome.
Start: L2 service desk at a Pune managed services provider, four years in, 5.8 LPA fixed, rotational nights, shell access to about 30 Linux VMs but no authority to change any of them.
Months 1 to 6 went roughly to plan. Linux lab built, VPC built by hand, SAA-C03 booked for month six. It was failed at 680 against the 720 pass mark, mostly on the cost and resilience domains, which is the common failure profile for people who learned from videos rather than from building. Retake passed in month eight, six weeks later than planned, at a second exam fee.
Months 9 to 11 lost time to a project crunch at work, so Terraform and CI slipped to months ten through thirteen. Applications started in month eleven and the first eighteen went unanswered, because the resume still opened with "Technical Support Engineer, incident management and ticket resolution" and screening tools never reached the projects at the bottom.
Rewriting the resume changed the response rate more than the certification did. Destination at month thirteen: junior cloud engineer at a different Pune MSP, 8.5 LPA fixed, day shift with a proper on-call rotation, and ownership of Terraform for two client accounts. Roughly a 45% move on paper, though about a third of that was simply being four years experienced and leaving an underpaying employer, not the cloud skills themselves.
Thirteen months, not ten. One failed exam, one crunch, eighteen ignored applications. That is what a normal version of this looks like.
Build the cloud engineer half of the switch on weekends, without leaving your support job
A 12 week live program preparing you for AWS Solutions Architect Associate (SAA-C03) and AWS DevOps Engineer Professional (DOP-C02), with hands on projects and mentor support. The current batch starts 27 Sept 2026 and classes run Saturday and Sunday evenings.
Explore the course
What Usually Goes Wrong
Transitions from support into cloud run long more often than they run short. Twelve months is the plan and fourteen to eighteen is the common outcome, because shift rotations and project crunches eat study weeks that never come back. A meaningful number of people land in a role that is support work with a cloud title and a 6% correction, then need another eighteen months to reach actual engineering. Some do not make it at all, usually because the study hours were never really available. None of this is a reason to skip the switch. It is a reason to plan for fourteen months and to keep your current job while you do it.
The certification-first trap
Passing SAA-C03 in month two feels like progress and produces nothing an interviewer can test. You will be asked to design a VPC on a whiteboard, and reciting service names is visibly different from having built one.
Costs you 3 monthsThe portfolio nobody can run
Screenshots of a console are not a portfolio. A repo with a README, a terraform apply that works on a clean account, and a note on what it costs per month is a portfolio. One of those is worth six of the other.
The resume that still reads support
If the first line says ticket resolution, screening never reaches your projects. Lead with what you built, keep the support role below it, and describe it in engineering verbs: automated, provisioned, reduced.
Costs you repliesThe internal move you never asked for
The cheapest route out is the cloud team two floors away, and most people never have the conversation. Ask in month three, not month twelve, so they can see you learning rather than arriving finished.
Costs you the easy pathThe Resume Lines and the Interview Answer That Move You
Three lines you can adapt today, assuming you did the work in the plan above. Copy the structure, not the numbers.
- Provisioned a multi-AZ AWS VPC (public and private subnets, NAT, bastion) in Terraform; environment rebuilds from an empty account in under 12 minutes with a single apply.
- Containerised a Python API and shipped it through a GitHub Actions pipeline with an automated rollback, tested by deliberately deploying a broken image.
- Handled 40+ P2 incidents a month against a 4 hour SLA across 30 Linux hosts; authored the runbook now used by the L1 queue for disk and service recovery.
That third line is your support experience written as engineering, and it is the one that makes the first two believable. Anybody can build a lab. Not everybody has been woken at 3am and made a correct call.
Then there is the question you will get in every single interview: why should we hire someone from support for an engineering role. Do not apologise, and do not oversell the lab. Answer in three beats: what you have already been trusted with in production, what you built to close the gap, and what you know you still lack. Saying "I have not run Kubernetes in production and I would expect to be slow on it for the first quarter" buys more credibility than any certification on the resume, because it tells the interviewer you can estimate your own limits, which is the actual skill they are hiring for.
One adjacency worth knowing, because support people are unusually well placed for it: if your environment is Microsoft-heavy, administering Copilot and agents is becoming its own role, and the Microsoft 365 Copilot and Agent Administrator course (AB-900) is a four week route rather than a twelve month one. It pays less than cloud engineering and it is a narrower ceiling, so I would only take it if you want out of shift work fast. If your interest bends toward data rather than infrastructure, the Microsoft Fabric data engineering program is the parallel path, and if it bends toward AI infrastructure, MLOps and agentic AI engineering sit one step beyond cloud rather than instead of it.
Also read: Docker Tutorial for Beginners 2026: Containerise a Python API in 7 Steps, which is the months 9 to 10 deliverable in the table above, written as a walkthrough.
What I Would Do in Your Position
Book nothing for the first eight weeks. Spend them on Linux and networking, because that is where the switch is actually won and because eight weeks is long enough to find out whether you like this work or just like the salary charts. If you still like it in week nine, pick one cloud, commit to the twelve month plan, and have the conversation with your own cloud team early rather than late. Keep the support job the entire time. The weekend-evening format exists precisely so that nobody has to resign to retrain, and resigning to study full time is the single most expensive mistake people make on this path.
If AWS is the platform your market runs on, the AWS Solutions Architect and DevOps Engineer course is where the SAA-C03 half of that plan gets built with an instructor in the room rather than alone at 1am.
Related guides
- DevOps Jobs in Dubai 2026: Salary Bands, Skills and Who Is Actually Hiring because the Gulf is where a lot of Indian cloud engineers make their second move.
- Data Engineer Jobs in Hyderabad 2026 for what the same GCC employers pay on the data side of the house.
- How to Become a Data Analyst Without Experience in India if you are weighing cloud against the shorter, cheaper analytics route.
- AI Engineer Jobs in Bangalore 2026 for where the cloud plus AI premium is currently concentrated.
- 8 Generative AI Project Ideas for 2026 if your portfolio needs one project that is not infrastructure.
Frequently asked questions
How long does it really take to go from IT support to cloud engineer in India?
Plan for 10 to 14 months at eight to ten hours a week, and expect the real number to land closer to fourteen once a shift rotation or a project crunch eats a few weeks. People who move internally into an existing cloud team sometimes start the new work sooner, but they typically take longer to be paid as an engineer.
Can I become a cloud engineer without a computer science degree in India?
Yes for most employers, with two caveats. Hyperscaler graduate programs such as the AWS Cloud Support Associate intake do filter on B.E, B.Tech, MCA or M.Sc degrees and specific graduation years. Managed service providers and startups, which is where most first cloud roles come from, screen on what you can demonstrate.
Am I too old to switch to cloud engineering at 35?
No, but the route changes. At 35 with ten years of support you are usually a poor fit for a junior cloud engineer opening and a good fit for a cloud operations or platform support role where your incident judgement is worth paying for. Target those, then move sideways into build work. Applying to fresher-shaped roles at that experience level is what makes the switch feel impossible.
Will I have to take a pay cut to move into cloud?
Often it is a flat move rather than a cut. First cloud-facing roles are commonly advertised in the 4 to 9 LPA band, so an L2 engineer already on 7 or 8 LPA will see offers at or slightly below their current fixed pay. A cut is worth considering only when the role hands you real infrastructure ownership, and almost never when it is support work with a cloud title.
Which should I do first, AWS SAA-C03 or Azure AZ-305?
Whichever platform your city and your current employer actually run. If you have no signal either way, AWS has more advertised Indian openings, with Naukri listing over 16,000 AWS cloud engineer vacancies when checked in September 2026. AZ-305 is cheaper from India at roughly ₹4,800 plus tax versus $150 for SAA-C03, and it suits Microsoft-heavy enterprises better. Do not do both in the same year.
Do I need Kubernetes to get my first cloud job?
Not for the first one. Entry-level cloud job descriptions and hyperscaler support intakes screen on Linux, networking and scripting. Kubernetes appears in mid-level cloud engineer listings alongside Terraform and observability, which means it is the thing to learn in your first cloud role rather than before it.
Is a home lab enough, or do I need production experience?
A lab is enough to get interviews if it is reproducible: a public repo, a README, Terraform that runs on a clean account, and a monthly cost figure. It is not enough to get through a senior technical round, which is why the support incidents on your resume matter. The two together are the case, not either alone.
What if my company has no cloud team to move into?
Then the internal route is closed and you are on the external one, which means the portfolio and the certification carry more weight and you should start applying in month eleven rather than month fourteen. Managed service providers are the most realistic first external employer, because they hire for potential more readily than product companies and they run enough client accounts to give you real infrastructure quickly.
About this guide. 360 Digital Transformation is an independent training provider offering exam preparation and skills training. We do not offer placement or job guarantees, and we are not affiliated with the certification bodies or employers mentioned. Salary figures are advertised market ranges collected from public job listings, not guarantees of what any individual will earn. Figures cited were checked on 12 September 2026.