Microsoft SQL Server

The Microsoft SQL Server Health and Microsoft SQL Server Activity checks monitor Microsoft SQL Server and complement each other. Microsoft SQL Server Health watches the slow-moving fundamentals — service state, connectivity, database states, backup age, disk space on SQL volumes, failed agent jobs, and Always On availability groups — and runs every 30 minutes by default. Microsoft SQL Server Activity watches live behavior — blocking sessions, severe error-log entries, and key performance counters — and runs every 5 minutes by default. Both run on Windows against SQL Server instances installed locally on the device.

Prerequisites

• You need the Monitoring Check Management permission. Your administrator assigns permissions under Administration → Groups.
• The device runs Windows with one or more local SQL Server instances, is online, and has the octoja agent installed.
• By default the checks connect with Windows authentication as the agent's system account. If that account cannot access your instances, create a monitoring login and enter it in the check — see Set up the database monitoring account.

Steps

1. Go to Devices and open the SQL server.
2. Click the Checks tab → Add Check → Microsoft SQL Server Health.
3. Configure what to monitor. The fields are described under Settings.

4. Click Add Check.
5. Add the second check the same way — Add Check → Microsoft SQL Server Activity.

Settings

Microsoft SQL Server Health

If a warning and critical value are in the wrong order, the check swaps them before evaluation. Always On availability groups have no field of their own: the check evaluates them automatically when the instance hosts replicas.

Microsoft SQL Server Activity

Results

Down or unreachable instances are alerted by the Health check only, so the two checks never page you twice for the same outage.

Microsoft SQL Server Health

Microsoft SQL Server Activity

Only blocking and error-log entries raise an alert. The performance counters (buffer cache hit ratio, page life expectancy, connections, blocked processes) are displayed but never raise an alert.

Tips

• Tune the backup-age thresholds to the customer's actual backup plan first — the defaults assume a nightly full backup and hourly log backups. Databases in the simple recovery model have no log backups, so they are only checked for their full backup. If a database reports missing log backups, it runs in the FULL or BULK_LOGGED model — set up log backups or change its recovery model.
• Use the instance and database filters to keep noise out of shared servers — for example, exclude a vendor's utility databases you do not manage.
• The error-log look-back window is deliberately short. If you raise it substantially, one old error can keep the check red across many runs even though the incident is over.

FAQ

Does disk-space monitoring warn before a SQL Express database reaches its size limit?

No. Monitor disk space on SQL volumes checks free space on the volumes holding SQL files, not database size relative to the installed SQL Server edition's limit. The check reports the edition but does not derive a capacity threshold from it. A database can reach its edition limit while the volume still has plenty of free space.

Monitor that limit separately using database-aware measurements for your SQL Server version and edition. A threshold on a single .mdf file is not an equivalent measurement of database capacity and should not replace it.