Background:
I'm on SAM 6.1.1, NPM 11.0.1 (I know I'm not on the latest. I'm planning on upgrading pretty soon)
We have been using the App Insight for SQL templates to monitor our SQL databases for about 2 years and have had no issues (generally) getting it set up and collecting data. Our SQL databases are set up in many different ways ( versions, clustering, etc.) but just recently two of the templates stopped polling on two servers that are in a SQL 2012 sp2 "always on" cluster The other servers in that cluster are polling OK.
I have tried/verified:
1. The account we're using "works" and can connect (via putting it in the template and running the tester returns "ok"). The SQL team has validated it has the rights it needs. The monitoring account we use is the same for all databases in this cluster. It is using windows authentication
2. I have tried removing and re-adding the AppInsight for SQL template to the servers. No change
3. I have tried removing and re-adding the SQL server to Orion (then the template). No change.
In addition: The SQL DBA's have indicated they don't see Orion even attempting to connect to SQL.
I have seen in the docs how for windows 2012 clusters you're not supposed to add the AppInsight for SQL template directly to each node in the cluster (which we have always done) but instead to attach it to the "cluster IP address". We have a 2012 active/active cluster and the data polling appeared to be working just fine (until it just stopped ). Is setting it to use cluster IP addresses actually necessary? Our DBA's questioned this because "it was working just fine all this time until it suddenly stopped" and I don't really have an answer as to "why" we would need to set it up like the docs suggest (short of collecting duplicate data and/or having junk left in the Orion DB when databases move around in the cluster, which I assume would be cleaned up during maintenance tasks (yeah, I know about the assuming...)
Also: I had the DBA's execute some of the troubleshooting steps I found to determine the cluster IP and the results are different that expected.
The SQL statement they ran:
SELECT SERVERPROPERTY('ServerName')
results in 2 server names (specifically, the two servers we are having issues with), not just one as the docs suggest it should.
Questions:
1. Are there any known issues or troubleshooting steps for APM or this template specifically? Telling me to "just do it like the docs say" is OK, but I'd like some reason why it needs to be done that way so I can convert all of our other similarly MS clustered databases over to that method.
2. If we DID convert the AppInsight for SQL template to just use the single cluster IP address (as the docs indicate) to monitor everything in the cluster...is Onion going to be able to handle it well? It's an 8 node SQL cluster with ~600 databases on it. We have 1 poller and NPM/SAM/etc. all installed on 1 server (Orion SQL DB is on a seperate server).
Thanks in advance!