Tag Archives: latency

AlwaysOn Availability Groups – Query to Find Latency Part 3

Like Part 1 and Part 2, I use a similar query to find latency in AlwaysOn Availability Groups. The queries will return one line for each DB in the group. Part 1 , we looked at a simple query to grab the current status of the replica. Part 2, we looked at populating a Global Temp table every minute to gain a historical perspective of our AG latency. This query will look into creating an email notification when certain criteria is met. It will send the current status to a specific set of email recipients.

The example will be around specific criteria set as a step used in a SQL Agent job. If you use something like this, you need to tweak it to you organization’s needs. I do suggest that you set the job to a frequency that will not overflow your inbox with notices.

Send an email from server when latency for a specfic database datbase is over x number of minutes.

  •  Alter If statement > value for specific needs. This example fires of fthe email when latency is over 60 minutes.
  •  Alter “database_id = DB_ID(‘MyAGDatabase’)” to specify datbase to monitor
  • Alter “@recipients=N’[email protected];” to specify recipients. I suggest using a email distribution list over listing spefic emails.

The example is pretty simple and intended for demonstration purposes.

Read more »

AlwaysOn Availability Groups – Query to Find Latency Part 2

Sometimes there is a need to maintain some historical content on how current your Availability Group is performing. The following script builds off the original script in Part 1. It populates a global temp table with information every minute. I really got the idea from Brent Ozar’s sp_SQL Blitz. I took the road of least intrusive by creating the ##LatencyCheck_tbl global temp table. Sergey Gigoyan has a good post on global temp tables and good links at the end of the post, too.

The ##LatencyCheck_tbl temp table is populated after 1 minute and every minute until cancelled. Need to make sue the table gets dropped after you have done you analysis. It is a good idea to keep an eye on the size of the table, too (execute sp_spaceused ‘##LatencyCheck_tbl’). Like any temp table, service restart will drop the table the process cancelled. If this is something you would like always run. you can look into creating a SQL Agent job or even a procedure that auto starts. I don’t suggest doing that. If you do choose to do so, you should make sure you monitor of have a process in place to reduce the size of the temp table. Also, look into grabbing Perfmon counters and Extended Events. Everything has overhead and we do not want to succumb to the “Watcher Effect”..DAH! DAH! [Organ music].

Read more »

AlwaysOn Availability Groups – Query to Find Latency Part 1

AlwaysOn Availability Groups latency can be a real concern. Microsoft provides a AlwaysON Dashboard to monitor and some administration of Availability Groups in SQL Server Management Studio(SSMS). Ben Snaidero as an excellent posting about the dashboard and T-SQL monitoring of AG.

The query below uses sys.dm_hadr_database_replica_states to query the current state. It can be run on the primary or secondary server. We use a readable secondary and it is nice to see how close to real-time updates. As a roll-your-own solution, you could create a SQL Agent job to monitor and send an email when certain criteria is met.

Read more »