Hello All,

Please note that this Monitor Pack contains Auto Remediation and if detects that any of the SQL Services are not running it will start them.  Please be aware that if you are running a Clustered Environment that you may want to have two Monitor Packs one for the Active Server with Auto Remediation and one without applied to the Standby Server or you may want to remove the Remediation completely.

Many Thanks for your attention. 

Updated with some minor Rule changes.

Hello Joesph

Does this SQL Pack also Support SQL 2008 R2?

This Monitor Pack should work ok with SQL 2008 R2.  If anyone has any issues with this pack on SQL 2008 R2 please let me know.

Kind Regards,

Hi!

What is the best way of adapting this pack to support monitoring of multiple MSSQL instances? From what I can see, this is not supported due to the structure of the performance counters used by the metrics. Are there any way of solving this without making custom metrics and rules for each named instance on every MS SQL server?

Thanks!

Best regards

Right now this is something that is not as quite dynamic as i would like in Monitor Solution.  To do this you simply need to clone the Metric and modify it for each Instance and then add it to a rule.

So no way to do this without making custom metrics, though this will be getting easier in future releases.

Kind Regards,

Great monitor pack Joseph!

I only got a problem, when i monitor a cluster SQL Server i always get one node with Critical errors, is it possible to integrate that with cluster service? To monitor node changes and alert if active node changed, but not to display a critical error if the other node is up.

Regards,

Daniel Oliveira

Hi Daniel,

You will need to use two policies for that, one policy for Active Node and one policy for Inactive Node and to do this you need to use the detection rules that will detect which server is the active server.

Best Regards,

Actually, I would not suggest a detection rule as the default interval for detection rules is 1 hour; meaning you might have an hour of inaccurate monitoring (of course you can change this).  I monitor several clusters (Exchange, SQL, and others) and simply use a powershell command to return a 1 or a 0, indicating if this node is active (script below).  Therefore, in my Monitoring Rule, I have two metrics defined, e.g.

1. If  “SQL Cluster” Group is active on this node
2. And If SQL Server Service is not running

powershell -noprofile -command "$x = Get-WmiObject -namespace root\mscluster -class MSCluster_NodeToActiveGroup | Where {$_.Partcomponent.substring(30, 11)-eq 'SQL Cluster'} | where {$_.groupcomponent.substring(21, 12) -eq (gc env:computername)} | measure-object; write-host $x.count"