It very much depends on what you're doing to be doing in Sage.  If you're just going to be importing GL entries then you can do it. If you were going to want to process 800k sales orders per month (200k per week or 25k per day) then you're going to want very strong infrastructure.

For one of our clients that, during busy season, can be processing 5-7 thousand AR invoices with credit card processing per day. We use IMan to pull the transactions into an intermediary database. Then we have Sage macros that IMan will trigger to process those transactions. It's all in a queue system so IMan can be importing new transactions while the macro is creating invoices/credits/debits in Sage at the same time. The macro will keep going creating 500 entry batches until there are no more unprocessed transactions in the intermediary database.

By separating the importing of the external data and the processing of that data we can take Sage down for whatever reason (like an upgrade) and when Sage is up it will just start processing transactions again.

I have a similar experience to Django.  In general it's fine.  In our case we were integrating 200,000 transactions a day into the GL as journals.  We used IMan to do the whole thing and it handled it perfectly well.  The only bottleneck was the speed of tyhe Sage API in processing so many transactions. Some optimisation was needed of course but all OK.