Last halving pump trxs took forever and were way too expensive. The only way it's going to continue up to anywhere near $1m is if scaling solutions are in place and functioning. If not, price will be slammed down again when faced with a failure to scale as before.
How do we know that those were real trxs and not just spam by Roger Ver, Jihan Wu and the usual suspects?
Obviously there was increased organic traffic happening and the mempool went up, but the blocksize was also obviously cluttered by bad actors. Tx batching was not in yet and segwit wasn't as well.
The question is: was the rise stopped because of tx fee being too high, or it crashed because it just had to crash as it went out of hand? Shall we risk a another hardfork drama to find out?