If you put ssh on another port, all of those automated login attempts disappear from your logs. The determined human attacker probably has port-scanned you and found ssh on another port, so you'll see just that attacker and can look into it more deeply if you feel the need. On port 22, you probably never bother to look at your failed-attempt log, because there's so much noise.
My question is basically: Do they exist? How is the work flow and how much time per year do they spend on it? What incident occurred and what value did they derive from the process. Was it cost-effective?
If I only got an email per week or so, I'd probably not filter them to a folder, and actually look at them.
[0] At work, other people deal with intrusion detection there, and ssh access is gated behind a VPN anyway.
But we realized that as attacks increased, it was pointless to look at who or how often we were being attacked, because it wasn't making our systems more secure. What did make them more secure was actually securing them: patching systems, implementing firewall rules, blocking or slowing down multiple login attempts.
Non-sophisticated attacks are trivial to prevent, and sophisticated attacks won't show up in logs. Everyone needs to stop logging ssh attempts and just secure their boxes and move on with life.
And that's why port knocking is stupid.
Just because you see or don't see attacks doesn't mean you are more or less secure. It just means you've finally noticed there is risk. It doesn't mean there is more or less risk. It's the same risk, it's just more visible.
Assume you move ssh to another port, say 23. In that case, anyone trying to connect to port 22 would find a closed port and the connection would fail, on a layer before login attempts get logged.