Microsoft SQL Server Monitoring
Performance and availability of production SQL Server databases, from the buffer pool to locks, with alarms when system properties exceed their thresholds.
An ideal starting point for troubleshooting
Microsoft SQL Server is the most widely used enterprise database solution on the Windows platform.
The Sentinet³ SQL-Server Pattern enables database administrators to monitor the performance and availability of production databases. Sentinet³ connects to the source database, monitors the column values in the system tables, collects the data and generates alarms when the database’s system properties exceed a given threshold.
| At a glance | |
|---|---|
| Pattern | Sentinet³ SQL-Server Pattern |
| Sources | Database system tables, Windows event log, SQL configuration |
| Scope | Server, buffer pool, page I/O, locks, indexes |
The Pattern’s checks
The Pattern includes 25 checks. Each has configurable thresholds and generates a notification when values exceed them.
Server
| Check | What it verifies |
|---|---|
| Reachability | The server responds on the network |
| CPU | Percentage of CPU utilisation |
| Memory | Memory usage |
| Disk | Free disk space |
SQL Server parameters
| Check | What it verifies |
|---|---|
| Buffer Cache Hit Ratio | Percentage of pages found in the buffer pool without reading from disk |
| Page Lookups / Sec | Requests per second to find a page in the buffer pool |
| Free Pages (Cumulative) | Available pages |
| Total Pages (Cumulative) | Total pages in the buffer pool, including database and stolen pages |
| Target Pages | Number of pages in the buffer pool |
| Database Pages | Pages with database content in the buffer pool |
| Stolen Pages | Pages used for other purposes, including the procedure cache |
| Lazy Writes / Sec | Buffers written per second by the Buffer Manager’s lazy writer |
| Readahead Pages / Sec | Pages read per second ahead of use |
| Page Reads / Sec | Physical database page reads per second |
| Checkpoint Pages / Sec | Pages flushed to disk per second by a checkpoint or other operations that flush dirty pages |
| Page Writes / Sec | Physical database page writes per second |
| Lock Requests / Sec | New locks and lock conversions requested per second from the lock manager |
| Lock Timeouts / Sec | Lock requests per second that timed out, including internal NOWAIT requests |
| Deadlocks / Sec | Lock requests per second that resulted in a deadlock |
| Lock Waits / Sec | Lock requests per second that forced the caller to wait |
| Lock Wait Average Time (ms) | Average wait time in milliseconds for each lock request that had to wait |
| Page Life Expectancy | Seconds a page stays in the buffer pool without references |
| Page Splits / Sec | Page splits per second caused by index page overflows |
| Full Scans / Sec | Unrestricted full scans per second |
| Time to connect to the database | Time to connect to the database |
Faults caught before they become unmanageable
- Events in the Windows log: Sentinet³ can monitor any SQL Server event reported in the Windows server’s event log.
- SQL configuration: combined with configuration monitoring, it makes it possible to catch specific faults before they become unmanageable, carry out impact analyses and track parameter behaviour precisely.
- Threshold alarms: when the database’s system properties exceed the set threshold, Sentinet³ reports it with an alarm.
Related topics
Want to see Sentinet³ on your infrastructure?
Tell us about your environment: sites, technologies, critical services. We will show you how Core and Infinity can work on your case.