When the evidence names the database, it says so. When it doesn’t, it stops.

Three databases have no recorded full backup. One backup job failed. Many tools would link the two and call it solved. This one linked the job to one database and said the cause for the other two is not determined.

I typed “full server check” on a noisy instance. It ranked the findings, each with a confidence score: three databases have never had a full backup, an enabled backup job failed, and four databases are offline.

I asked about backups. The agent also ran the job checker, said so, and kept the two tools’ findings separate. The failed job’s own error names a backup file:

Cannot open backup device ‘C:\SQLBackupsArchive\FULL\SalesOrders_FULL.bak’. Operating system error 3 (The system cannot find the path specified.). BACKUP DATABASE is terminating abnormally.

So the agent tied the failed job to SalesOrders. For HRPayroll and FinanceArchive, it said the cause is not determined.

Then I asked it directly. It said yes: the failed job is identified as the SalesOrders full-backup job, and SalesOrders has no recorded full backup. It then drew its own boundary. The data does not establish whether this job failure accounts for the entire backup history gap.

What I want from an on-call tool

When the evidence names the database, it says so. When it doesn’t, it stops. A wrong causal claim at 3am costs far more than a DBA making an obvious inference, and most tools in this space guess at that point

Ran on Azure OpenAI (gpt-6-sol), AgentDBA v1.13.2

Want to see it on your own estate? www.agentdba.ai