Gabriele Santomaggio

Ok I can reproduce the issue with three nodes cluster and this configuration: ```csharp var config = new StreamSystemConfig() { UserName = "test", Password = "test", Endpoints = new EndPoint[]...

Oh ! that's an interesting topic. There is also [Istio Traffic Management](https://istio.io/latest/docs/concepts/traffic-management/), [here](https://github.com/Gsantomaggio/k8s/blob/wip/rabbitmq_traffic/istio-configuration/rabbitmq-istio.yaml#L42-L48) an example with RabbitMQ. The same with HA-Proxy, we can also play with the TCP Backlog queues...

Just a thought ... Would be interesting to try to implement an [`XDP_DROP`](https://prototype-kernel.readthedocs.io/en/latest/networking/XDP/implementation/xdp_actions.html ) It would drop / avoid ddos attach or not valid connections [interesting post to read](https://blog.cloudflare.com/how-to-drop-10-million-packets/)

>Can we agree on K8S as the baseline? > Ingress makes this much easier to capture, follow along & keep it relevant for longer-term. I'd like to add that the...

Ok, I did some test with Traefik using SNI and strict sni configuration, using a configuration like this: ``` apiVersion: traefik.containo.us/v1alpha1 kind: IngressRouteTCP metadata: name: rabbitmq-big-ingress spec: routes: - match:...

Thank you @Hawxy for your comment. Feel free to open a PR to improve it.

closed via https://github.com/rabbitmq/rabbitmq-stream-dotnet-client/pull/159

@ricsiLT Thanks for the PR! I didn't check the code, but there must be a bug somewhere. a lot of connections are recreated This is why the assert `Assert.Equal(1, connectionsKilled);`...

@ricsiLT Thank you again for your contribution. I close this PR in favour of #163 !

Can't find anything. There is also: - https://github.com/force-net/Crc32.NET since we will need some compress library there is also: - https://github.com/icsharpcode/SharpZipLib