I have web application (NodeJS) and I plan to deploy it to AWS. To minimize the cost it will run on single EC2 instance. I'm worried though about what will happen if someone decides to bless me with DDOS attack and hence have few questions. Now, I did quite a bit of research, but as my understanding is clearly lacking I apologise if some of the questions are plain stupid:
I want to avoid people flooding my site with layer 4 attacks. Would it be sufficient to set my Security Group to accept traffic only (in additions to SSH port 22) from:
Port Range 80
Would above stop UDP flood and others from hitting my EC2 instance?
Via Security Group I would allow SSH connections to port 22 only from my static IP address. Would that keep attackers away from trying to attack port 22 completely?
My EC2 instance will run Ubuntu. I want to avoid application layer attacks (layer 7) and was planning to do it directly from my application, so somehow detect if certain IP floods particular URLs and block them if necessary. This however seems a bit late as the traffic already hits my web server and my server have to do the work anyway. So instead of doing this directly from my application I was thinking if that was possible to use IP tables to block any dodgy traffic before it comes to my web server. Is there set of some common settings that would be able to recognise rogue behaviour and block offenders? I was planning to look into fail2ban in hope this would simplify the process. Now, I do understand if it gets that far it will hit my EC2 instance anyway, but I want to protect my application also from e.g. brute force attacks.
Would AWS CloudFront take care of most DDOS Layer 4 attacks? If not then using free CloudFlare would make any difference?
Say that someone floods my website anyway and this results in more traffic then I anticipated. Is there any way to stop charges at some point? There are billing alerts but I cannot see any way to set hard limits on AWS and say get instance offline if bandwidth exceeded.
I also do realise that there is now way to completely prevent DDOS attacks but I want to protect at least against basic attempts. Thank you in advance for any help.