Cloud cost optimization (FinOps)
A lower cloud bill in weeks, without anything your customers would notice.
For teams whose Azure, AWS or Kubernetes bill grows faster than the business. I find where the money goes, cut what nobody needs in the order that saves the most with the least risk, and leave a routine behind so the savings stay.
Sounds familiar?
- The monthly invoice keeps rising and nobody can say which team or service caused it
- Kubernetes nodes run half empty because requests were set by guesswork
- Dev and test environments run all night and every weekend
- Reservations were bought for machines that no longer exist
How I work on it
- 1
See the cost
I pull amortized cost per resource from the billing API, tag every line with product, team and stage, and build a run-rate forecast so a change shows up in days, not at month end.
- 2
Check before cutting
For every candidate I check live what runs on it, who uses it and the peak load, never the average. Anything a customer could notice stays untouched until it is proven safe.
- 3
Right-size and switch off
Kubernetes requests follow real p95 usage so the autoscaler can remove nodes, idle resources go, and non-production runs 60 instead of 168 hours a week.
- 4
Fix the expensive parts
Costly queries, debug logging in paid storage and data kept longer than anyone needs. In one case a single query fix cut the load on Cosmos DB about 40×, with identical output.
- 5
Commit last, then keep it there
Reservations and savings plans only for the new, smaller footprint. Then budgets with forecast alerts, a daily cost report in your team chat and a monthly review keep the bill from creeping back.
Interactive
How many nodes do your requests really need?
Move the sliders to your numbers. Requests decide where pods fit, so oversized requests buy nodes nobody uses.
Node size
5
nodes today
2
nodes after right-sizing (200m)
60%
less node capacity
CPU only, about 90% of a node is allocatable, at least 2 nodes for availability. Check memory the same way before you remove a node.
All calculators and checklistsWhat you get
- Cost baseline per resource, team and environment with a run-rate forecast
- A ranked list of changes, each with impact, risk and the way back
- The changes themselves, one ticket each, rolled out without customer impact
- Budgets, forecast alerts and a daily cost report in Teams, Slack, Google Chat or email
- A one-page monthly report for management with a two-month forecast
Questions buyers ask
How long until we see savings?
The baseline takes a few days. The first changes usually land in the second week, and with a run-rate forecast you see their effect within days. In the engagement shown here, total Azure costs were 48% lower after five weeks.
Will our customers notice anything?
No. Every change is checked live against peak load first and comes with its way back. Anything a customer could notice stays out of scope until it is proven safe.
Does it work on AWS and Google Cloud too?
Yes. The method is the same on Azure, AWS and Google Cloud, and for Kubernetes on any of them. Only the billing APIs and the names of the discounts differ.
What do you need from us?
Read access to billing and monitoring, one person who knows the platform, and a short weekly slot to decide on the next changes. Write access only where we agree on a change.
How do you price it?
After a 30-minute call: either a fixed scope with a defined result, or a few days per month as an ongoing FinOps lead. No surprises, the scope is written down before we start.
Can my team keep it running afterwards?
Yes, that is the goal. The reports, alerts and the checklist stay with you, documented, and I walk your team through the monthly review until they run it alone.
Read more
Get in touch
Tell me what you're building. I'll reply within 1–2 business days.
- Company
- MADO Consulting
- Follow me on LinkedIn
Available worldwide, remote. On-site on request, depending on the country.
Opens Calendly's booking calendar here. Loading it sends data to Calendly (see Impressum).