SQL Server Estate Assessment

SQL Server 2016 left extended support on July 15, 2026. SQL Server 2017 follows on October 13, 2027.

If you are running either one, you are in one of three positions. You are paying Microsoft for Extended Security Updates, you are running without security patches, or you are somewhere in between and not certain which.

Most of the teams we talk to are in the third position. Not because nobody was paying attention, but because the estate grew over fifteen years and the inventory lives in four different places, none of which agree.

Dates as published on Microsoft's product lifecycle pages.

Start with the inventory. It costs nothing.

Before you buy anything from anyone, find out what you actually have.

We will send you a read-only script. It collects version, edition, patch level, database count and size from each instance and writes the results to a file. It reads system views. It does not change configuration, it does not read your data, and you can read every line of it before you run it. Your DBA runs it, not us.

Send us the output and we will read it and get on a call to walk through what it shows. No charge, and nothing owed afterward.

For a fair number of organizations that call is the whole engagement. You find out you have four instances instead of seven, two of them were decommissioned in 2019 and never came off the list, and the one nobody worries about is the one still on 2016. Now you have something you can put in front of your CFO.

What you are carrying while it sits there

The version does not stop working on the end date. Nothing breaks. That is the problem, because nothing happening is easy to keep not noticing.

Security patches stop unless you buy them. Microsoft sells Extended Security Updates as a paid add-on, and once extended support ends that is the only route to a patch. Vulnerabilities found after the end date are not fixed on instances without it.

The paid route expires too. Extended Security Updates for SQL Server 2016 run three years and end July 17, 2029. There is no year four. An instance still on 2016 after that date is permanently unpatched at any price. Working backward from 2029 through procurement, testing and a change window is how this becomes a 2027 budget line rather than a 2028 emergency.

Bug fixes stop, not just security fixes. Out of support means no cumulative updates at all. A functional defect you hit in 2028 is one you work around permanently.

Auditors ask. Unsupported software carrying known unpatched vulnerabilities is hard to defend in a HIPAA risk analysis, a SOC 2 review, or a client security questionnaire. It tends to come back as a finding with a remediation date attached, and at that point the timeline stops being yours.

Insurers ask. Cyber liability carriers have started asking about end-of-life software at renewal. It can affect what you pay and what is covered.

Your application vendors stop certifying. Software vendors drop out-of-support database platforms from their support matrix on their own schedule, not yours. You can end up unable to take a required application upgrade because the database underneath it is too old.

It rarely travels alone. An instance still on SQL Server 2016 is usually sitting on a Windows Server version with its own end-of-support date, on hardware with its own replacement cycle. Those costs tend to arrive together, which is an argument for finding out now rather than next year.

One more thing worth knowing before you plan a migration. SQL Server 2019, which a lot of people still treat as current, left mainstream support on March 1, 2025. It is in extended support until January 9, 2030, which means security fixes only. No new functionality and no non-security bug fixes. Landing a 2026 migration on 2019 buys you less runway than it looks like.

When the inventory is not enough

Sometimes it is, and we will tell you so. Sometimes the answer is that you have a real problem and you need a plan with numbers attached, because a budget request without a defensible estimate does not survive the meeting.

That is what the assessment is for.

What you get:

• A complete inventory across every instance, including the ones that are not on anybody's list
• The support status of each one, and what it costs you now versus what it costs next year if it does not move
• A risk ranking, so you know which instance to move first and why
• A migration sequence that accounts for what depends on what
• A written cost estimate for each move, in a form you can hand to Finance without translating it first

If you engage us for the migration work within 90 days, the full $8,500 comes off the project price. The assessment is not how we make money. It is how we find out whether there is a project worth doing.

Where we have done this

We work with health plans, provider groups, TPAs and clearinghouses, mostly in Florida. We have done this migration many times in environments with SSIS, replication, Always On, and the kind of undocumented dependency that only announces itself at cutover.

If your estate looks different from that, the work is the same. The version matrix does not care what industry you are in.