Persistent periodic spike in CPU% utilization

I have been hosting my web api server on Digital Ocean for over 2 years now. In last 5-6 months I noticed my droplet dashboard has a persistent periodic (generally every 1 hour) CPU% utilization spike as shown in this image link . At first I thought that may be I accidentally pushed some orphan process via my git pull in the api which is running periodically. So I started to investigate my code. However I found that there is nothing wrong with my application. So my suspicion increased and then I started checking my authentication logs. There I saw huge number of malicious SSH auth attempts and subsequent failure logs. Although I have UFW and Fail2Ban in place for protection and I am using passwordless ssh keys, but these spam events cause spikes in my server cpu usage, disk IOs and network in outs. I am just wondering if this is normal (which I do not think because a few months ago this was not that frequent). I have also noticed a dramatic upsurge in my web server 4xx logs. Is anyone else experiencing such events. If yes, then what counter measures have you taken. If no, then can anyone please guide me how such events are mitigated?