mehrdadrad / tcpprobe Goto Github PK

The tcpprobe seems to work well in the k3s cluster. It probes some sites continuously and I can see tcp-related metrics in the Prometheus and Grafana.
I follow the config file described in https://github.com/mehrdadrad/tcpprobe/wiki/helm , where i set a nginx deployment and annotate it to tell tcpprobe to "probe" it.
But after that I deployed the Nginx server. I found several things confusing.

• the annotation tcpprobe/targets: "http://PODIP:80" didn't get rewritten , which means that PODIP is literally PODIP and so i know the Nginx won't be "probed"
• even if i fill the right ip address of the nginx server using another pod's annotation, the Nginx server still didn't get probed. I see no data in the metrics it expose.

Env:
OS X + k3d cluster (k3s in docker)

I would appreciate all the help, thanks!

apiVersion: apps/v1 
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 1
  template:
    metadata:
      labels:
        app: nginx
      annotations:
        tcpprobe/targets: "http://PODIP:80"
        tcpprobe/interval: "10s"
        tcpprobe/labels: '{"app": "nginx"}'
    spec:
      containers:
        - name: nginx
          image: nginx:latest
          ports:
            - containerPort: 80

Looks like a number of values are not available when running docker on ARM or maybe rasberrypi specifically?

• RTO
• RcvMss
• SndMss
• pmtu
• RTT
• Congestion window size

Is this a platform limitation? I used the default docker-compose file when I cloned the repo on both x86 and arm, and the arm one doesn't display these values.

Considering the amount of stats the tool provides is there any recommended way one can use the tool to estimate the bandwidth of a link by sending probes to some remote site?