We have two databases, one is 10TB in size and one is very highly transactional. Both have been dismissed as candidates for high-availability groups because of these qualities. The way I see it is as follows:
a. The only problem with a 10TB database in a HAG is the amount of space required in the initial build of the HAG when adding the database to the group. This space is only temporary and although a bit of a bind, still needs to be taken into consideration when designing a backup plan. That makes this a non-problem and is not a reason not to include a database of this size in a HAG.
b. A very highly transactional database in a HAG is only a problem when the database runs in synchronous mode with high latency between the endpoints. This is because every commit has to be acknowledged from the secondary before it can be committed in the primary. This leads to performance problems in databases with a high level of transactionality. To resolve this issue the database should be made highly available within the group asynchronously so that acknowledgements aren't required by the primary before they can be committed. This has the disadvantage of the possibility of lost records in the event of a failover but for two servers that as are close together as ours this possibility is so small as to be negligible.
These are the arguments that I will present in order to remove the worries and doubts our admins have with regards to putting these into HAGs.
Does anyone have any thoughts or ideas, maybe something I have missed or misunderstood?