Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/nirmata/kube-netc
A Kubernetes eBPF network monitor
https://github.com/nirmata/kube-netc
ebpf kubernetes networking prometheus
Last synced: 3 months ago
JSON representation
A Kubernetes eBPF network monitor
- Host: GitHub
- URL: https://github.com/nirmata/kube-netc
- Owner: nirmata
- License: apache-2.0
- Archived: true
- Created: 2020-03-26T20:50:33.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2021-01-26T23:18:08.000Z (almost 4 years ago)
- Last Synced: 2024-06-18T22:56:12.907Z (7 months ago)
- Topics: ebpf, kubernetes, networking, prometheus
- Language: Go
- Homepage:
- Size: 562 KB
- Stars: 80
- Watchers: 7
- Forks: 16
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# kube-netc: A Kubernetes eBPF network monitor
[](https://travis-ci.org/nirmata/kube-netc) [](https://goreportcard.com/report/github.com/nirmata/kube-netc)
kube-netc (pronounced kube-net-see) is a Kubernetes network monitor built using eBPF
## Getting Started
To test the current capabilities of **kube-netc**, this guide will walk you through viewing the network statistics of your nodes.
### Install kube-netc
First, install the daemon set using the install.yaml:
```
kubectl apply -f https://github.com/nirmata/kube-netc/raw/master/config/install.yaml
```
### View results
This will start the **kube-netc** DaemonSet on your cluster and setup the required roles. Then, we get the name of the kube-netc pod:
```
kubectl get pods --all-namespaces | grep kube-netc
```
For example, my **kube-netc** pod is:
```
kube-netc-j56cx
```
In a new terminal, we port-forward the port of our pod so we can access it with *curl* outside the cluster with:
```
kubectl port-forward kube-netc-j56cx 9655:9655
```
9655 is the port we are going to access the Prometheus endpoint on. We can then curl the /metrics endpoint using curl on local host to show the Prometheus metrics:
```
curl localhost:9655/metrics | grep bytes_recv{
```
This is an example output of the query showing the total bytes received by this node from each given connection:
```
...
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:1640",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 2960
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:38256",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 295276
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:38258",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 178
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:38446",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 276120
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:38448",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 39315
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:38460",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 122540
bytes_recv{component="kube-controller-manager",destination_address="172.18.0.2:38496",destination_kind="pod",destination_name="kube-controller-manager-drewcluster-control-plane",destination_namespace="kube-system",destination_node="drewcluster-control-plane",instance="",managed_by="",name="",part_of="",source_address="172.18.0.2",source_kind="pod",source_name="kube-controller-manager-drewcluster-control-plane",source_namespace="kube-system",source_node="drewcluster-control-plane",version=""} 3382
...
```
As we see the bytes received by each connection is shown and the source IP is given. If there is a known pod, node or service with the same IP, the *source_name* and or *destination_name* is also given.
## Grafana Demo
There is a pre-prepared Grafana dashboard so you can test out **kube-netc** yourself and visualize the reported stats.
Make sure to point Prometheus to the **kube-netc** exporter in your configuration by appending a new job to your *scrape_configs* in your *prometheus.yml*:
```
- job_name: 'kube-netc'
static_configs:
- targets: ['localhost:9655']
```
Then you can start Prometheus and import [the dashboard](grafana/grafana_kube_netc_demo.json) into Grafana.
## Design
Please see the [DESIGN](DESIGN.md) for information on how kube-netc is structured.
## Thanks
Special thanks to Alban from Kinvolk for his assistance with implementing eBPF and DataDog for their eBPF library.