What I automate / Cloud cost and housekeeping
The cloud bill nobody can explain
Cloud spend grows quietly. Nothing is obviously wasteful, and yet the bill is up again, and the honest answer to why is that nobody has the time to find out.
Most of the waste is boring and mechanical: environments running at nights and weekends when nobody is working, disks and snapshots left behind by things that were deleted, oversized instances from a load test that finished last year, and resources with no tag so nobody knows who to ask.
This is among the easiest work to justify, because the saving arrives on a bill you can compare month to month. It is also the work that most often pays for itself before the invoice for it is due.
You probably know this is you if
- The bill grows and no single person can explain the increase
- Non-production environments run 24 hours a day
- Nobody can say which team or project a given resource belongs to
- A certificate or domain has expired outside working hours at least once
- There are snapshots and volumes attached to nothing
The jobs in this area
A bill that grows and nobody can explain
A monthly breakdown by team, project or tag, with the biggest movers since last month named.
Test machines left running over the weekend
Non-production resources stopped on a schedule and started when people need them.
Nobody knowing what a resource is for
Tagging enforced at creation, and a list of the untagged things that already exist.
A certificate or domain expiring on a Sunday
Expiry dates watched, with warnings early enough to act during working hours.
Old snapshots and disks quietly billing forever
Unattached and orphaned resources reported, then cleaned up on rules you approve.
How a build like this goes
- 1
Measure before changing anything: a breakdown by team, project or tag, and the biggest movers since last month. Half the value is in the visibility.
- 2
Then the safe, reversible wins: schedules that stop non-production out of hours, and a cleanup of orphaned resources on rules you approve rather than rules I choose.
- 3
Then tagging enforced at creation, plus a list of the untagged things that already exist, so the problem stops growing while you work through the backlog.
- 4
Then the expiry watch — certificates, domains, reserved commitments — with warnings early enough to act during working hours.
A worked example
Weekend shutdown for non-production
- Before
- Three environments running continuously, used only in office hours.
- After
- Stopped at 9pm, started at 8am on weekdays, off entirely at weekends. A start button for anyone who needs them outside those hours.
Office hours are about 50 of the 168 hours in a week. Stopping non-production outside them removes roughly 70% of their compute cost, and compute is usually the largest line. On a ₹40,000/month non-production spend that is around ₹28,000 a month against a quick-win build — it pays for itself in the first or second month, which is why I usually suggest it first.
Those are illustrative figures, not a quote and not a promise. Your own numbers are the only ones that matter — the calculator does the same arithmetic on them.
What this normally costs
A job in this area is usually a quick win: ₹18,000 to ₹45,000, ready in 3 to 7 working days. The exact number comes from the audit, in writing, before anything is built.
The audit itself is ₹9,999 and comes off the first build in full. If it turns out this is not worth automating for you, that is what the report will say.
Questions about this kind of work
Will this break something?
Schedules and cleanups are built to be reversible and are applied to non-production first, on rules you approve. Nothing is deleted on my judgement alone.
We are on a reserved or committed plan.
Then the savings are different and often smaller, and I will tell you that before quoting rather than after. Visibility and expiry-watching still apply.
Can you just tell us what to do instead of building it?
Yes — that is an audit, and for smaller estates it is often the right and cheaper answer. Several clients have implemented it themselves afterwards, which is a perfectly good outcome.
Tell me about your version of this
Tell me the job you are tired of
One call, half an hour. You will get a straight answer about whether it is worth automating — including 'no' if that is the answer.