Free 20-point SQL Server health check
A SQL Server health check is a structured review of an instance against a fixed set of configuration, performance, backup and security checks. Onsys runs 20 of them free of charge on one Australian SQL Server instance, using read-only access, and gives you the written findings whether or not you engage us afterwards.
The 20 points, in full
Every check below runs against a live instance using read-only access. You get the findings whether or not you engage us afterwards.
Backup coverage
Every database is checked for a current full backup, and for log backups where the recovery model requires them. Databases with no backup at all are the single most common finding.
Restore verification
We look for evidence that a restore has actually been tested. An untested backup is a hypothesis, not a recovery plan.
Recovery model vs stated RPO
Simple recovery caps your worst-case data loss at the last full backup. We check the model matches the loss you have actually agreed to tolerate.
Corruption checks
When DBCC CHECKDB last ran cleanly on each database, and whether it is scheduled. Corruption found late is often corruption already inside every backup you hold.
Version and patch level
Build number against the Microsoft support lifecycle, so you know which instances are past end of support or missing security updates.
Instance memory configuration
Max server memory against physical RAM. Left at the default, SQL Server will contend with the operating system for memory.
Parallelism settings
MAXDOP and cost threshold for parallelism against your core count and workload type — still at defaults on most estates we review.
tempdb configuration
File count, equal sizing and autogrowth. Misconfigured tempdb is one of the cheapest large performance wins available.
Data and log file growth
Autogrowth increments, percentage-based growth and unconstrained log files that can fill a volume without warning.
Index health
Fragmentation levels, duplicate and unused indexes, and the write cost they impose on every insert.
Missing index pressure
What the optimiser has been asking for, weighted by impact — treated as evidence to review, never applied blindly.
Statistics currency
How stale the statistics are and whether auto-update is keeping pace with the volume of change on the largest tables.
Wait statistics
The instance top waits, which tell you what the server is actually spending its time waiting for rather than what you assume it is.
Blocking and long-running queries
Current blocking chains and the queries with the worst cumulative duration, reads and CPU.
Deadlock frequency
Deadlocks recorded in the system health session, with the pattern behind the repeat offenders.
Storage latency
Read and write stall figures per file, which separate a slow query problem from a slow disk problem.
SQL Agent job health
Failed and long-running jobs, jobs with no failure notification configured, and jobs still owned by departed staff accounts.
Privileged access review
Who holds sysadmin, which logins are orphaned, and which accounts have privileges nobody can now justify.
Authentication and encryption
Authentication mode, the state of the sa account, and whether TDE and connection encryption are in place where the data classification calls for it.
High availability readiness
Always On, failover clustering, mirroring or log shipping configuration — and whether a failover has ever been tested.
What you receive
A written report you own, whether or not you become a client.
- A findings report rating each of the 20 points, with the evidence behind every rating.
- A prioritised remediation list separating what is urgent from what is merely untidy.
- An estimate of effort for each remediation item so you can budget or schedule it internally.
- A 30-minute walkthrough call with the senior DBA who ran the check.
- The raw query output, so your own team can reproduce every finding.
- No obligation to engage Onsys for the remediation — plenty of clients do it themselves.
How we work with you
Book a time
Pick a 30-minute slot. We confirm which instance you want reviewed and what your change process requires.
Grant read-only access
A read-only login for the duration, or a screen-share session if your policy does not permit external logins at all.
We run the checks
About two hours against a live instance. The queries are read-only and safe to run during business hours.
You get the report
Within three business days, followed by a walkthrough call with the DBA who ran it.
Book your free 20-point SQL Server health check
One instance, no charge, no obligation. Most clients book the walkthrough call within a week of the report.
Book the free health checkPublished
Common questions
Yes — one instance, no charge, no obligation and no requirement to engage us afterwards. You keep the report and the raw query output either way. We offer it because roughly half the environments we review turn out to have a finding serious enough that the owner wants help fixing it, and the other half tell someone we were straight with them.
About two hours of read-only checks against the live instance, which can run during business hours without affecting production. The written report follows within three business days, and we walk you through it on a 30-minute call.
A read-only login with VIEW SERVER STATE and db_datareader on the system databases is enough for every one of the 20 checks. If your policy does not allow external logins at all, we run the same checks over a screen-share while a member of your team drives. No data leaves your environment.
A findings report rating each of the 20 points with the evidence behind it, a prioritised remediation list with effort estimates, the raw query output so your team can reproduce every finding, and a 30-minute walkthrough with the DBA who ran the check.
No. Every check is a read-only query against system views and dynamic management views. Nothing is written, no configuration is changed, and no maintenance window is required.
Yes, though the free offer covers SQL Server. Oracle, PostgreSQL, EDB and MySQL health checks are run as a fixed-price engagement — usually a half day per instance at $150 per hour, quoted before the work starts.
No. The check is run remotely, so we cover organisations anywhere in Australia and New Zealand. Our consultants are Melbourne-based and work Australian business hours, with 24/7 cover available on a support plan.