Topology graph from NetObserv showing kube cluster external traffic to AWS eu-west-3 S3
📢 Just published: Identifying #kubernetes cluster external traffic with subnet labels and #netobserv
netobserv.io/posts/identi...
#network #observability #ebpf #kubernetes #openshift
(Some inaccuracies for netobserv but hey, that's ok 😊)
www.tigera.io/blog/kuberne...
#Kubernetes, 4 network observability solutions compared: Tigera Calico, Cilium (Hubble), Microsoft Retina and Red Hat NetObserv.
You can work on the #eBPF cache timeout/size to increase aggregation in time, which generates more or less flows.
We're changing the defaults. Some tests have shown UP TO 40% CPU DECREASE across all components. It's also less flows to be ingested by Loki.
Change: github.com/netobserv/ne...
NetObserv 1.11 for #openshift will be released in a couple of days.
FYI, if you use NetObserv already: there's something you might want to look at, without waiting.
In some scenarios, tweaking config can result in HUGE performance improvements.
1/2
ovn-kubernetes v1.1 is out today!
Check it out: github.com/orgs/ovn-kub...
Nice milestone for the k8s CNI, with cool features like network segmention (UDN) and BGP support.
NetObserv 1.9 already supports UDN and we're working to add more observability for BGP soon!
#kubernetes #ovn #bgp
In our last blog, Mohamed explains how netobserv uses #ebpf to track #IPsec usage (and potential errors) within a #kubernetes cluster. Check it out!
👇
netobserv.io/posts/monito...
We just released netobserv 1.9! (community #k8s and #openshift downstream)
Featuring IPsec tracking via #ebpf, more ingest-time flows filtering capabilities, some CLI improvements, performance improved and some bug fixes.
Release note: docs.redhat.com/en/documenta...
(upstream == downstream)
Netobserv (Red Hat) is #hiring a software engineer!
Check out the job description: redhat.wd5.myworkdayjobs.com/fr-CA/jobs/j...
Oh, that's a nice and long post on #kubevirt written for #vSphere admins who might consider migrating:
veducate.co.uk/kubevirt-for...
As said, #redhat is well placed here. For those admins looking into #openshift #virtualization , NetObserv can help monitor closely your workloads network.
One more thing, that isn't mentionned in the above article: the recent versions of ovn-kubernetes have a new feature called UDN for User Defined Network, which is basically segmentation. More about that here: developers.redhat.com/articles/202...
#redhat #openshift #kubevirt #ovn #kubernetes
Oh, that's a nice and long post on #kubevirt written for #vSphere admins who might consider migrating:
veducate.co.uk/kubevirt-for...
As said, #redhat is well placed here. For those admins looking into #openshift #virtualization , NetObserv can help monitor closely your workloads network.
Interesting post from #netflix: netflixtechblog.com/how-netflix-... on #eBPF based #observability with network flows.
If you're not Netflix but still are interested, you can also take a look at what we're doing with NetObserv: netobserv.io/about/
- OVN is improving its observability, for instance it can emit events such as when a network policy blocks/accept traffic. NetObserv can leverage and display that.
It's only Tech Preview at the moment, so you'll need to wait a bit before having it on a prod cluster.
There are also cool new features related to ovn-kubernetes (you won't get them if you use a different CNI) :
- As #OpenShift introduces new User Defined Networks (docs.openshift.com/container-pl...), which offers a much better isolation, NetObserv can visualize them as topologies and more...
... and more filtering options in the #eBPF agent. For example, you can now decide to sample ALL external traffic while maintaining a 1:50 ratio on internal traffic, to save resources without compromising what matters most.
More on that here: netobserv.io/posts/enhanc...
There is also a good bunch of improvements on the CLI side, such as generating Prometheus metrics on the fly: netobserv.io/posts/networ...
One of the cool things that you can immediately notice if you upgrade from 1.7, is the performance improvement, especially on the eBPF Agent CPU. More on this here: netobserv.io/posts/perfor...
NetObserv 1.8 is out!
Check it out: netobserv.io/posts/what-s...
Available as a helm chart or, for #OpenShift users, in OperatorHub.
We also created a CNCF slack channel, come and have a look if you want to chat: cloud-native.slack.com/archives/C08...
#kubernetes #network #observability
Joël présentant NetObserv
On continue avec Joël qui nous parle eBPF et présente son projet @netobserv.bsky.social
📢 New blog post on how to correlate pods and services traffic in #Kubernetes : netobserv.io/posts/enhanc...
This feature will be coming soon, in netobserv 1.8 (already merged in the main branch).
It's achieved by hooking into the #linux #kernel conntracking, with #ebpf .
Good morning Bluesky !
#Introduction
This is a developer account of the NetObserv tool. It's an #opensource software, born at #redhat, that provides network observability through #ebpf probes in the Linux kernel. It's mostly designed for #Kubernetes / #OpenShift
Our blog: netobserv.io
