Five Questions Your Azure Should Answer Every Week

1. Who touched what?
   Are the people acting in your subscription allowed to be there? Or are there forgotten accounts, ex-contractors, and service principals nobody remembers creating?

   
   

2. Did the people acting have the proper training?
   Compliance frameworks require it. Auditors ask about it. Nobody tracks it.

   
   

3. Which resources cost you the most?
   Not the total invoice - the top ten line items, ranked, in front of the person who signs the check.

   
   

4. Who is touching those specific resources?
   For each of the top ten, who provisioned them, who last scaled them, who owns the answer to "is this still needed?"

   
   

5. Any resource trending up?
   A database that cost 200 EUR last month and 800 EUR this month is either a growing success or a runaway you will notice in three months when it is 8,000 EUR.

Your cloud has the answers. All of them. Right now. What is missing is the weekly review that turns data into decisions.

What we found on our own infrastructure

We built our review service for ourselves first and ran it against our own production. Some of what it surfaced:

• Two production sites had been failing their daily backups for five days.
  Nothing alerted us. If we had needed to restore, we would have discovered it the hard way.

• Certificate renewals failing daily for weeks
  because DNS points to a CDN instead of the origin. Not urgent - the CDN handles it - but daily noise that nobody had questioned.

• Silent config debt
  on resources we had forgotten we owned.

None of these were exotic. All were sitting quietly in place the whole time. The information was there. The attention was not.

And the finding that started this

Some months ago, during a routine review at a client, we noticed a database configured for ~20,000 USD per month. Nobody had questioned the bill. The tier was wrong for the actual workload - probably scaled up during an incident that had long passed, then never scaled back down. It had become part of the invoice's baseline. Just another line item nobody looked at.

That finding paid for months of work on its own. But it was an accident. Luck is not a strategy.

The gap is not tools. It is ritual.

Every enterprise cloud already exposes cost, activity, access, and posture data. Most publish recommendations continuously. None of it costs extra. What is missing is the ritual - the weekly hour when someone reviews the right findings and asks the right person "is this still needed?" Nobody's job description says that, so nobody does it consistently. That is not a failing. It is how every enterprise looks.

Our service closes that gap. Not another dashboard - a weekly review that surfaces what deserves attention, attributes it to the person who knows, and remembers the answer so nobody has to ask twice.

Why CFO, tech lead, and compliance officer all care

For the CFO: every silent cost anomaly is money paid for nothing. A single 20K/month finding pays for years of review discipline. And the service itself is a fixed, small, predictable line item.

For the tech lead: every silent failure - a backup that has not run, a cert that cannot renew, a resource nobody remembers - is a future incident with your name on it. The service catches these while they are still cheap to fix, without adding to your team's workload.

For the compliance officer: every unowned access, every untrained user with production rights, every unreviewed change is an audit finding waiting to happen. The service produces the evidence trail that ISO 27001, SOC 2, and PCI DSS all require - as a byproduct of doing something useful anyway.

Three different jobs, one shared ritual, one small monthly cost.

What we are offering

A weekly review service that runs against your cloud subscription and answers those five questions - with the specific resource, the specific person, and a suggested action - delivered to the people who can act on them. Findings and decisions are stored so the same question does not come back next month.

Small, predictable, easy to try. If it catches one anomaly per year, it pays for itself many times over. If it catches nothing, you sleep better knowing so.