Availability Group Threads

My understanding is that on the primary that the minimum pool size would be (assuming all dBS are active)

• 1 Message Handler Thread per AG
• 1 Log Scanner thread  per database
• 1 Log Send thread per database * # of secondary replicas

Is this correct?

Looking at Bob Dorr's diagram  ( https://blogs.msdn.microsoft.com/psssql/2012/05/17/alwayson-hadron-learning-series-worker-pool-usage-for-hadron-enabled-databases/ ) it appears that each secondary would need at least 2 * the number of databases + 1 or more message handler threads. 

Is that correct?

thanks