Good to know
No. It is built to support DBAs and Oracle teams, not replace them. Complex, high-risk, or ambiguous situations are explicitly pointed toward a human specialist rather than answered with false confidence.
Where possible, guidance is checked against real infrastructure behavior and documented Oracle context — reproduced on a matching environment rather than left as an unverified, fluent-sounding guess.
Yes. Students and newer DBAs use the tools to safely explore real ORA errors and Oracle internals without the risk of touching a production system.
A generic chatbot can produce a confident-sounding answer with no way to check it. ORAcraft starts from Oracle-specific context and applies validation logic where possible, and is upfront about the difference between the two.
Guidance spans Oracle 11g through 19c, 21c, and 23ai, across on-premises, RAC, Data Guard, and the major cloud platforms.
Good to know
The senior DBA who scoped your engagement is the one who delivers it. Every engagement is led and signed off by a senior Oracle DBA — 10+ years in production — and never delegated to a junior team.
Fixed scope, fixed quote, agreed before any work starts. No open-ended hourly clock and no surprise invoices partway through an engagement.
Not cold. Every engagement is scoped to get in, do the work, and hand your team a system they can run without us. Once we’ve delivered and it’s running well, we’re happy to talk about staying on to keep running it ourselves — a different model from embedding one of our DBAs under your direction, below.
Yes — that’s our second engagement model. A senior Oracle DBA from our vetted bench joins your team, on-site or remote, part-time or full-time, working under your direction. It’s different from managed services: you run the environment and direct the work; we supply the senior capacity and stand behind its quality with a named ORAcraft lead. Availability is confirmed at scoping — senior people are never in infinite supply, and we won’t pretend otherwise.
Yes — we regularly deliver as a specialist under a prime vendor or integrator, on their paper or ours. Professional indemnity insurance, senior sign-off, and documented handover apply either way.
Tell the assistant what’s going on. A few quick questions get you straight to the specialist who’d handle it — verdicts on your exact problem come from a person, not a bot.
No. Discovery work is read-only by default, and anything that changes production requires your written sign-off first.
Safe Passage Assessment — Oracle 19c → 26ai
Read-only access to your 19c estate, and about two hours of your DBA’s time. Nothing is executed and nothing is changed. Your DBA sits in the session and watches every command — or runs them themselves, if they’d rather. We don’t mind. Nothing happens without someone from your side watching it happen.
They can run the checks. What they can’t do is evaluate their own work from the outside — they built the workarounds and applied the patches, so the things that are invisible to them are invisible for a reason. It isn’t a competence problem, it’s a geometry problem. And when the board asks who verified it, ‘we did’ is a weaker answer than a signed report from someone with no stake in the outcome.
Then that’s what the report says, and you have the evidence to defend it. We’re paid for the verdict, not the migration — we have no interest in telling you to move. A ‘not this window’ verdict with reproduced evidence behind it is worth more than a green light nobody checked.
Then you have a signed document saying so, and you walk into the cutover knowing what you’re walking into. That is exactly what you’re buying: not a list of problems, but a number you can put your name on. ‘We looked, here’s what we found, and here’s what we couldn’t test’ is the answer to the only question that matters afterwards.
A health check tells you what your monitoring already told you. We reproduce. Every critical finding is run on a matching Oracle environment before it’s marked verified — and every report states, explicitly, what we could not test and why. Most audits don’t have that section. It’s the section that makes the rest of the report worth reading.
Database Platform Feasibility Assessment
PostgreSQL, SQL Server, and managed cloud flavours such as Aurora and RDS — and staying on Oracle is always priced as a candidate alongside them. The exact candidate list is fixed at scoping, based on your constraints and your team’s skills — not on what anyone prefers to sell.
Then that’s what the report says, and you have the measured case to defend it. We’re paid for the answer, not the migration — a ‘stay’ verdict backed by counted objects and timed workloads protects your budget exactly as much as a ‘go’.
Not as part of this assessment — deliberately. The assessment is priced and delivered standalone, so the verdict never feeds a migration quote. If the verdict is ‘go’, the roadmap works with whoever executes: your team, your integrator, or — as a separate conversation on separate terms — ours.
The free analysis is presales — it exists to start a migration, and it’s priced accordingly. Ours is the product itself: findings reproduced on a replica, the limits of every test stated in writing, and the cost of staying measured alongside the cost of leaving. Nobody selling you a move includes that last number.
Discovery is read-only, and your DBA watches every command — or runs them. Reproduction happens on a replica, never on production. And one rule holds for the whole engagement: you can stop it at any point, and everything verified up to that moment is documented and yours.
Estate Health Review
No. The report is yours, with no follow-on obligation of any kind. Most people who take the review are not buying anything yet, and a fair number of them never do — because the verdict was that their team is fine and they just needed the noise turned down.
It is optional before Watch, which is advisory — we tell you what we see, your team decides and executes, so the risk stays with you. It is required before Operate and Assure. On those we take responsibility for your backups, your restores and your incidents, and we will not sign a service level for an estate we have not inspected. If a standby has been silently out of sync for two years, that becomes our problem the day we sign, and yours the day it matters.
It is deducted in full from your first monthly invoice. On Operate at €6,200 you pay €3,300 in month one, then €6,200 from month two. It is a credit, not an inclusion — if you decide against a plan you keep the report and owe nothing further. The credit applies per estate, not per client.
Then we say so, in writing, and hand you the alert tuning on the way out. It has happened before and it will happen again. We would rather tell you that than sell you cover you do not need — and you would rather hear it from someone who just looked than from someone who wants a monthly invoice.
Read-only access to the estate, about two hours of your DBA’s time, and one agreed window for the restore test. Nothing is executed in production without someone from your side watching it happen — your DBA sits in the session and watches every command, or runs them themselves if they would rather.
Two things. We perform a restore rather than reviewing a backup log, and we give a verdict rather than a list of observations — including "manage it in-house" or "hire someone", when that is what the evidence supports. A health check tells you what it saw. This tells you what to do about it.
A senior Oracle DBA with more than ten years in production — the same person who would own your estate if you went on to a plan. Not a junior with a checklist, and not an account manager reading someone else’s findings back to you.
We start from the count, then check the weight. Forrester’s benchmark for large enterprises is roughly forty databases per DBA, but the real constraint is volume — around five terabytes per person. One DBA can carry twenty-five 200 GB databases or five one-terabyte ones, so the count tells us where to start, not where to land. At the review we confirm the band against actual size: many small schemas move you down, a handful of very large transactional databases move you up — we say so before you sign, not after. Estates above 100 databases are on request, not configured.
Open-source stack deployed — Zabbix, Prometheus, Grafana — or integrated with Oracle Enterprise Manager, Datadog, or whatever you already pay for. Commercial tooling is bought by you, directly, at your own rate. Never marked up, and no commission is taken from any vendor, ever. We are an independent Oracle practice, not a reseller — there is no version of this where our recommendation pays us twice.
Blog
No fixed schedule and no filler to hit one. A new article goes up when we’ve actually reproduced something worth writing down.
The same senior, certified Oracle DBAs who run our assessments and services — never a ghostwriter, never AI-generated copy dressed up as a technical note.
Yes — tell us the ORA code, the symptom, or the decision you’re stuck on through the contact form, and we’ll consider it for a future article.
No. The assistant answers your specific, live situation. The blog is general-purpose — patterns and methods that apply broadly, not a substitute for a grounded answer on your own environment.