Best Practices

SQL Best Practices

•Separate the Operating System from the database files and the log files.

•Separate the database files from the log files.

•For database files, performance increases with more spindles included in your RAID configuration.

•Use SATA with TCQ support or SCSI Drives, the faster the RPM the better.

•For better recoverability, use a SCSI interface instead of SATA and IDE.

•For large bandwidth demands on the I/O bus, use a different bus for the transaction log files.

•If there is a DBA on staff, defer to the DBA.

If using SQL Express, increase the number of "Keep nn databases" in ELM Database Settings, on the Archive tab.  If data volume nears SQL Express database size limits, ELM will automatically rollover primary or failover databases.  These will be counted, along with archive databases, toward the "Keep databases" limit.

Database Best Practices

Database Settings - Event data can quickly fill a database. Plan to keep only the data you need, for as long as you need it. Microsoft SQL Server 2008 R2 Express has a 10GB limit.

Recommendation (based on monitoring 10 or more servers):

•Keep informational, audit success, and warning events for 7 days. Add a prune specification to the Database Settings in order to remove event data older than 7 days.

•Keep error events for two weeks, then prune them.

•Keep audit failure events for one month, then prune them.

•If physical disks is available, separating the Windows swapfile, SQL .ldf file, and .mdf file onto separate physical disks can help overall performance.

Recommendation (for compliance purposes):

•SQL Server Standard R2 or Enterprise R2 editions are preferred over SQL 2008 R2 Express Edition.

•Keep events in the ELM primary database for two weeks, and then archive events. Using a shorter time period will improve performance if the number of events is extremely high.

•Automate archiving the Archive Database. You should expect to have several multi-gigabyte archive database files. These files may be moved to removable media as prescribed by your compliance plan.

•Configure Performance Data Collectors to aggregate data weekly, and delete annually. This will provide one week detail history, and 52 weeks of summary.

Monitor Item Best Practices

•Only Service Agents can execute Monitor Items in real-time. For Virtual Agents, we recommend a Scheduled Interval of 10 seconds or greater.

•Some Monitor Items include the ability to execute Actions. Leverage this capability for additional management power!

•When using Virtual Agents, the Monitor Items are executed remotely by the ELM Server. You can use the ELM Server performance object to obtain metrics about the Monitor Item job queue.

Notification Method Best Practices

•Set Threshold settings in order to reduce the impact of event storms.

•Select the best notification method. Use the ELM Advisor desktop notification for critical events such as Errors and Audit Failures.

•When using e-mail Notification Methods for events you review as part of a daily routine, but do not necessarily need to know about immediately, use an Exchange Public folder as the destination.