I have a SQL 2016 (SP1) 3 node AG and I am having some performance concerns that lead me to believe that using indirect checkpoints instead of automatic may be beneficial. It is a large memory, very active primary and I just want to make sure I mitigate risk and don't want to add performance issues, even if temporarily, due to the criticality of this server consuming crucial data points.
Are there any caveats to switching this setting on a live, very active DB?
I currently don't have a maintenance window on this particular AG (something I'm working on with management)
*This DB isn't already using the default setting of 60 for Target Recovery Time because the database was moved and upgraded and not created under 2016.