Some things are worth the wait. We're back. The quest continues

Expert service

Is your SQL Server slow — or just badly configured?

After two decades of tuning production SQL Servers, we can tell you: it's almost never the hardware. The same ten traps account for most of the slow queries, 3 AM timeouts, and "we probably need a bigger server" meetings we get called into — and most are fixed in days, not quarters.

The ten traps we find over and over

1. Factory defaults in production MAXDOP untouched, cost threshold for parallelism stuck at 5, unlimited max memory, a "Balanced" power plan — settings tuned for a desktop, quietly throttling your ERP.
2. Missing indexes — and dozens of useless ones Queries scan millions of rows to return hundreds, while duplicate and unused indexes tax every single write.
3. Parameter sniffing The same stored procedure runs in 40 ms for one customer and 4 minutes for another, because the cached plan was built for somebody else's data.
4. Implicit conversions An nvarchar parameter meets a varchar column and your carefully built index gets scanned end to end — no error, no warning, just heat.
5. Stale statistics The optimizer estimates 3 rows, receives 3 million, and builds a plan for the wrong universe.
6. TempDB contention Every sort, spill, and temp table on the instance queuing on the same allocation pages of one overworked file.
7. Blocking chains & NOLOCK band-aids One long transaction stacks readers up behind writers; sprinkling NOLOCK everywhere trades the timeouts for silently wrong results.
8. Row-by-row processing Cursors, WHILE loops, and scalar functions grinding through in 20 minutes what a set-based query returns in 2 seconds.
9. Autogrowth at the worst moment Files growing 10% at a time and a log shredded into thousands of VLFs: the database stops to grow itself exactly when it's busiest.
10. Buying hardware to fix code Doubling the RAM to outrun a table scan buys you months. Fixing the query buys you years.

The fix: a SQL Server system review

We run lightweight, read-only diagnostics against your server — nothing installed, nothing changed — then walk your team through a short, prioritized plan: what's hurting you, what to fix first, what to ignore, and what each fix should buy you. No finger-pointing, no jargon, no 400-page report.

Remote-friendly  ·  typically scheduled within a week  ·  you keep the plan whether or not we do the work