I have at least half their subnets blacklisted at this point. They seem to host a lot of bot traffic, port scans, and other generally unsavoury characters.
Is this the wrong approach? A losing battle of whack-a-mole?
FWIW I get a not-insignificant amount of malicious traffic from AWS, Azure, and Google but I view these providers as "too big to block" - I can't blacklist large swaths of their IP space without breaking the Internet.
- Only keep open ports/forward ports for applications you use, drop/block everything else.
- Use strict host-header checking for web services on port 80/443, drop anything to 403/404 that doesn't have a valid host-header for the website(s) you're hosting.
- Move SSH and other remote admin servers to use a non-standard port. (legit, find a random port number between 9000-65535)
- If it doesn't need to be public, allow-list it with iptables.
Unfortunately DO and other providers will never have 100% legit traffic, it's just the nature of the Internet's noise floor.
Hope this helps you or someone else!