AppInsightUser for SQL cluster monitoring with dispersed instances

sankeyj

I am trying to apply the AppInsightUser for SQL template to some clustered servers hosting our SQL instances.   I followed the instructions provided in the guide, added the virtual IP of the cluster, selected the instances on the node (virtual IP), assigned the port and credential to each instance. 

When I added the virtual IP of the cluster it resolved to the name of one of the member servers.  The name it populated was not necessarily the server the active instance was on, where does it pull that name from?   I have both member servers in SolarWinds, but am only monitoring basic health and status.  I do not have any of the instances selected in their resource list.

There are five instances on the cluster, they are not all on the same server, they are dispersed between the two.  Each instance has an IP address different than the host.   I don't think SolarWinds can tell which server is active because of this.   None of the instances will authenticate through the AppInsightUser monitor.   Neither the server or the database show any login failures or attempts.

From the SolarWinds server (database manager) we are able to authenticate and connect directly with the IP address of the instance.  So we know the credential works with the instance and the network is passing the traffic.

Our theory is that because the virtual IP address is not the same as the instance, the monitor cannot connect with the instance itself.  That it is trying to connect to the server.    I tried to add the IP address of the instance itself and SolarWinds doesn't know what to do with it because it just wants to connect via WMI.

So can I monitor clustered SQL instances that:

1.  Do not have the same IP address of the host server

2.  Are not all active on the same member server

3.  Is there a way to set up the ApplnsightUser monitor to poll the instance by IP address?

Thank you,

Jan

sankeyj

I did a work around for my problem.  I monitor the cluster IP of the two members, but only monitor the server portion.   I do not apply any of the instances on it.   I do select any shared network mounts and shares since they float between the two member servers.

I added each instance individually, using it's unique IP address. Initially I added them with ICMP, but it would always show the node being down because we don't allow ICMP traffic.  The better solution to gather more detailed information was to allow WMI traffic between the IP addresses of the instances and the SolarWinds server.  Also, without WMI, you won't get all the statistics in the AppInsightUser template to poll.  Each instance has an domain account with the appropriate permissions to respond to the AppInsightUser account I have on the SW server.  The firewall rules also contain the specific port for each instance as they are statically assigned.  Since the instances in the cluster will failover to the other server, I only need to monitor the instance IP address and not worry about which member server it's on. 

I should note, when I would add an instance, the name would automatically change to the member server it was currently on. I would edit the name back to the instance name to avoid confusion.

Hope this helps someone else.