EC2

check back in cost explorer in February

commands

sudo yum update
sudo amazon-linux-extras install -y nginx1 && sudo systemctl enable nginx.service && sudo systemctl start nginx.service start nginx
aws --region=us-east-1 ec2 describe-spot-price-history --instance-types t4g.micro --start-time=$(date +%s) --product-descriptions="Linux/UNIX" --query 'SpotPriceHistory[*].{az:AvailabilityZone, price:SpotPrice}'
aws ec2 describe-images --owners amazon --filters "Name=name,Values=al2*" "Name=architecture,Values=arm64" --query 'sort_by(Images, &CreationDate)[-1].ImageId' --output text

spicy

echo 2 > /proc/sys/vm/drop_caches requires being sudo
- this freed 85MB / 20% of memory, could be cron every 12 hours if you want to make Linus unhappy. I noticed after an hour it went back to its regular value of 126M / 34% usage
echo vm.vfs_cache_pressure=1000 >> /etc/sysctl.conf requires being sudo
sudo sysctl -w vm.swappiness=10
sudo sysctl -w vm.vfs_cache_pressure=200

AL2 Arm t4g.nano has 169MB free, 79MB used / 435MB. (0MB used in swap, if this is high then it needs to be expanded)
in theory socket.io can handle ~1,000 (200MB) concurrent connections on this machine
redis should only be used on machines with at least 4GB RAM, t4g medium would cost $7