The unique problem · The Stability Trap
A stable Oracle system doesn’t mean a safe one — it means nothing has forced its hidden risk to surface yet. A migration is that moment, all at once.
A house can look perfect for 30 years — until you open the walls and find wiring nobody logged. Your Oracle 19c estate is that house: every undocumented patch and deprecated feature sits quietly, waiting for cutover night.
Your system’s stability doesn’t measure how ready you are to migrate. It measures how long nobody has looked.
What your monitoring sees
— beneath the surface —
Why teams skip the readiness check
Three things teams tell themselves before a 26ai migration — and why each one is the trap, not the shortcut.
Internal team
They built the workarounds and applied the emergency patches. You can’t objectively evaluate from the inside what’s only visible from outside — it’s geometry, not competence. Surgeons don’t operate on themselves.
Monitoring
Monitoring catches active failures, not dormant incompatibilities. A deprecated feature throws no error in 19c — it just stops in 26ai. A smoke detector, not a building inspection.
Fix as we go
Reactive works when downtime is cheap. On a mission-critical cutover, every hour of unplanned outage runs into six figures. "We’ll fix it as we go" isn’t a strategy — it’s a gamble.
What we do instead
Every critical finding is reproduced by a senior DBA on a live Oracle environment before it’s marked verified — with the limits of each test stated honestly. Always verified. Never assumed.
The stakes
The business gave you a maintenance window. Six hours, on a Saturday, agreed months in advance. Everything else is negotiable. That number isn't. And the only honest way to know whether the migration fits inside it is to have already run it — on a replica of your production data, and timed it.
‘It should take about four hours’ is a sentence with no evidence behind it. Nobody knows how long your datapump export takes until someone runs it against your data volume. Nobody knows how long your indexes take to rebuild until someone rebuilds them. Until then, the number in the change request is a guess wearing a suit.
If step 7 of 12 runs two hours long, you don’t get two extra hours. You get a decision, at 4am, with the business waiting: push through and finish late, or roll back and explain why. Both are the same conversation on Monday. The difference is whether you saw it coming.
A migration doesn’t fail because Oracle is hard. It fails because something surfaced that nobody knew was there — an object that wouldn’t compile, a feature that no longer exists, a dependency nobody documented. Every one of those is findable on a replica, in advance, on a Tuesday afternoon, with nobody waiting.
What we do instead
That's the whole job: turn every unknown into a measured number, before the window opens. Not because the migration is dangerous — but because signing off on one you haven't measured is.
Product · 01
A signed go/no-go verdict on your 26ai migration, backed by evidence — plus a plan your team can execute without cutover-night surprises.
Scope: in-place upgrade · 19c estates of any size · rollback points named
Every assessment is led and signed off by a senior Oracle DBA — 10+ years in production. The verdict is never delegated.
What you'll get — five deliverables
Select a deliverable to exploreTap a deliverable to explore
What we do
We inventory dependencies, deprecated features, patch history and licensing exposure across your 19c estate — read-only. Nothing is touched in production without your written sign-off. You receive a documented map of what’s actually there, not what the docs claim.
What this will feel like
We’ll ask for read-only access to your system, and about two hours of your DBA’s time. That’s the moment most people hesitate — so: your DBA sits in the session and watches every command. If they’d rather run the commands themselves, they run them. We don’t mind. Nothing is executed without someone from your side watching it happen. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
Each gap is verified at the right level — static / CPAT checks where those settle it, reproduction on a matching Oracle environment where behaviour is the real question. Every finding states the method used and what it does and doesn’t tell you about your production, with prioritised risk scores and a clear go/no-go verdict.
What this will feel like
You will see gaps you didn’t know existed, and some of them will have been sitting there for years. Reading that list is uncomfortable. It is better here than on cutover night. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
A step-by-step plan your team can execute: sequenced tasks, flagged dependencies, rollback points defined where they exist — and the points where rollback is no longer clean named explicitly, so nothing surprises anyone on cutover day.
What this will feel like
Some rollback points don’t exist. We write that down, in your plan, in plain words: past this step, there is no clean way back. That line is hard to read. It is also the line that stops you from discovering it at 3am with the phone ringing. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
A non-technical summary of findings, risks and recommended action — built so you can justify the spend and carry the decision internally without being in every room.
What this will feel like
The verdict may not be the one you wanted. Sometimes the honest answer is not this window, or stay where you are. If that is what the evidence says, that is what the deck says — and you will have the evidence to defend it in the room. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
A scope-and-definition session at the start: we agree what "verified" means for your specific infrastructure, which findings warrant reproduction and which are settled by static checks. No surprises about scope or deliverables.
What this will feel like
We’ll ask you to define what verified means for your environment before we start. You may find your own team doesn’t agree on it. That conversation is uncomfortable. It is also the one that decides whether everything that follows is worth anything. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
We inventory dependencies, deprecated features, patch history and licensing exposure across your 19c estate — read-only. Nothing is touched in production without your written sign-off. You receive a documented map of what’s actually there, not what the docs claim.
What this will feel like
We’ll ask for read-only access to your system, and about two hours of your DBA’s time. That’s the moment most people hesitate — so: your DBA sits in the session and watches every command. If they’d rather run the commands themselves, they run them. We don’t mind. Nothing is executed without someone from your side watching it happen. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
We inventory dependencies, deprecated features, patch history and licensing exposure across your 19c estate — read-only. Nothing is touched in production without your written sign-off. You receive a documented map of what’s actually there, not what the docs claim.
What this will feel like
We’ll ask for read-only access to your system, and about two hours of your DBA’s time. That’s the moment most people hesitate — so: your DBA sits in the session and watches every command. If they’d rather run the commands themselves, they run them. We don’t mind. Nothing is executed without someone from your side watching it happen. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
Each gap is verified at the right level — static / CPAT checks where those settle it, reproduction on a matching Oracle environment where behaviour is the real question. Every finding states the method used and what it does and doesn’t tell you about your production, with prioritised risk scores and a clear go/no-go verdict.
What this will feel like
You will see gaps you didn’t know existed, and some of them will have been sitting there for years. Reading that list is uncomfortable. It is better here than on cutover night. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
A step-by-step plan your team can execute: sequenced tasks, flagged dependencies, rollback points defined where they exist — and the points where rollback is no longer clean named explicitly, so nothing surprises anyone on cutover day.
What this will feel like
Some rollback points don’t exist. We write that down, in your plan, in plain words: past this step, there is no clean way back. That line is hard to read. It is also the line that stops you from discovering it at 3am with the phone ringing. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
A non-technical summary of findings, risks and recommended action — built so you can justify the spend and carry the decision internally without being in every room.
What this will feel like
The verdict may not be the one you wanted. Sometimes the honest answer is not this window, or stay where you are. If that is what the evidence says, that is what the deck says — and you will have the evidence to defend it in the room. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
What we do
A scope-and-definition session at the start: we agree what "verified" means for your specific infrastructure, which findings warrant reproduction and which are settled by static checks. No surprises about scope or deliverables.
What this will feel like
We’ll ask you to define what verified means for your environment before we start. You may find your own team doesn’t agree on it. That conversation is uncomfortable. It is also the one that decides whether everything that follows is worth anything. And one rule holds for the entire engagement: you can stop it at any point, for any reason. Everything verified up to that moment is documented — and it’s yours.
You can stop here. Fixed scope and fixed price, in writing, before this starts. Nothing runs until you approve it.
How the engagement runs
Read-only discovery.
Your DBA sees every query we run. Production is never touched.
Reproduction on a replica.
Findings become evidence instead of opinions — with a short note after each session.
The verdict session.
Go/no-go, the evidence behind it, and the roadmap — walked through live, with your team in the room.
Timeline is set at scoping, in writing, with the quote — it scales with your estate, not with our calendar.
Get started
60 minutes. Free. No pitch. You talk to the senior DBA who would run your assessment — not a salesperson.
What to bring
Your Oracle version(s), rough estate size, and your deadline. That’s enough.
What happens on the call
We ask about your migration window, your known risks, and what ‘verified’ needs to mean in your environment. You ask anything — including how we’d handle your specific setup.
What you leave with
A straight answer on whether this assessment fits your situation — including ‘it doesn’t.’ If it doesn’t, we say so on the call and you’ve lost an hour, not a budget.
Prefer email? hello@oracraft.com
Request sent
We’ll be in touch shortly to schedule your scoping call. hello@oracraft.com.