Main metrics to monitor with Always On

Question

There are some metrics to monitor for performance with Always On: redo rate, received rate, send queue,..etc What are the main ones that we should usually monitor and keep attention to?

Answer 1

There are plenty of DMVs that you could use to monitor your AG. Which metrics you want to monitor and the threshold you give to them depends on you and your environment. For some, as long as the data is moving that's all they care about. Other environments may want to be alerted when things start slowing down. Plus, there are a ton of ways and use cases to configure and utilize Availability Groups. But you should at least have some alerts set up. I'd set up some alerts to start with... with these at a minimum.

• Error 35264 for Suspended Data Movement
• Error 35265 for Resumed Data Movement

Then, you may want to look into some others. You can find them in sys.messages. I added a where clause to show a range where these are grouped together, but they could be elsewhere in sys.messages.

SELECT message_id [error_number], severity, text
FROM sys.messages 
WHERE message_id between 35254 and 35269 
AND language_id = 1033

Brent has a good list of other alerts you should have created too.

More on Alerts via the Docs