Network

Add an option in the configuration so the load balancer listens on an ip on the private VPC, therefore request made to it via private ip range will be forwarded to the destination also on the private ip range.

It would be great to have the possibility to set "SameSite=None" for Load Balancers with Sticky sessions enabled. It's necessary when using iframes in a cross domain situation.

Since TLS has support for Server Name Indication, there's no reason the load-balancers shouldn't be able to handle terminating multiple certificates. The alternative is to pass-through the SSL to the droplet, but this then means the load-balancer can't inject in the clients true IP with X-Forwarded-For.

So far, load balancers only get an IPv4 and no IPv6. In practice, this means that load balancers cannot be used if you want to provide IPv6 connectivity. I would love to use this feature, but only if IPv6 was made available.

I want to be able to have a load balancer between App Platform apps, and potentially even any IP address (including non-DigitalOcean managed) or domain name. The purpose for this in my case is to make use of the Global Load Balancer, across the same App Platform apps in different regions, as well as in other cloud services by IP or domain name - potentially as multi-cloud redundancy, or partial migration of an application to DigitalOcean.

Recently UDP port support was added to Load Balancers (see also e.g. https://docs.digitalocean.com/products/kubernetes/how-to/configure-load-balancers/#protocol ). This allows for various useful applications, in particular support for HTTP/3. However, this protocol does not allow using PROXY Protocol as of yet (as has been confirmed by my contact with DO support). This means that it is not possible to use the client IP for any purpose when using both HTTP/3 and TLS Passthrough on a Load Balancer. This idea suggests implementing the PROXY Protocol for UDP traffic as well, so that the original client IP is exposed to the service behind the Load Balancer.

We can add firewall to Load Balancers through API, and it requires to provide not just the IPs to whitelist but all the required fields for a load balancer, which is already set up through UI, and it is inconvenient. It would be perfect if we could set firewall rules straight from Load Balancer Settings along with the other settings!

not only droplets but also load balancers should be put behind the firewalls

This functionality will allow you to dynamically change the number of replicas, depending on the number of current connections. Since the available number of simultaneous connections per 1 replica is 10000, this is more than enough for normal use. But when the number of connections increases instantly, for example, during a DDOS attack, the limit of 10000 simultaneous connections is reached, which leads to the fact that all resources become unavailable due to the limit being reached. To solve the problem, you can keep more than 10 replicas at all times, for example, 10, but the cost of LoadBalancer will be increased 10 times, which is not efficient from the financial side.

I have a need to support rate limits by source IP for example for inbound traffic to my load balancer. Can we add this feature?