ebpf_exporter VS cilium

There's also the ebpf_exporter which has modules to let you get per-device IO latency data.

For example, we alert when the database starts to get slow, because there's real user impact to slow database queries. We track disk IO latency, but it's for debugging after we get alerted by database alerts. Because database problems could be so many different things. Query plans changing, CPU saturation, etc. Slow reads/writes is just one cause of database problems.

There also a prometheus exporter for the data if you have prom set up. https://github.com/cloudflare/ebpf_exporter

If you have Linux, there is the ebpf_exporter. We use it to monitor real-time disk IO latency, with basically no overhead to the system.

My current go-to to start with storage monitoring is the ebpf_exporter. It allows you to wrap the underlying kernel functions with timing wrappers, which allows you to gather detailed data on every IO that happens. Gather that up into a histogram, and ship it to a monitoring system for visualization.

They would have had to add a few externals to get to Graduated but it's definitely a minority:

https://github.com/cilium/cilium/blob/main/MAINTAINERS.md

Next-gen networking thanks to Cilium

Hello there, I recently have the need to proxy my pod traffic through WireGuard. I initially have my eyes on https://github.com/angelnu/pod-gateway but I just couldn't get it working. It turns out that Cilium made a CVE patch couple years ago that basically nuked ability to do inter-pod encapsulated traffic (https://github.com/cilium/cilium/issues/15991). I wonder if there is any other way that can let me do this without switching out of Cilium? Thank you guys in advance :)

I have used Cilium as CNI and installing it with helm.

I'm using the default RaspiOS Lite 64bits and as highlighted in this issue, the RaspiOS kernel does not support CONFIG_ARM64_VA_BITS_48, which makes cilium-envoy to fail building. As solution, I was told to use either Ubuntu as base OS or Traefik Ingress Controller, which is not configured in K3s.

Working on integrating cilium and loxilb as a hobby k8s project. Both are eBPF based and will be interesting to see what will be the final outcome.

Particularly in Cilium, Gateway API is very proof-of-concept. So much so that you can't even change the type of the underlying service (or anything else about the generated object) yet.

Authentication using mTLS was later merged into cilium (https://github.com/cilium/cilium/pull/24263). It uses mTLS between cilium agents to authorize flows, but do note that the mTLS auth is de-coupled from the datapath transport (i.e. you need to configure cilium to use ipsec or wireguard, as otherwise traffic won't be encrypted). As a consequence, there are some gaps in the implementation right now, like packet drops. see https://github.com/cilium/cilium/issues/23808

network plugin to be used, based on the documentation. (Project Calico ,Flannel, Cilium )