traefik ingress not forwarding tcp messages

Question

I am trying to aggregate raspberry pi logs (IoT Devices) into Logstash/ElasticSearch running in EKS.

filebeat is already running in EKS to aggregate container logs.

This is my manifest file

apiVersion: v1
kind: ConfigMap
metadata:
  name: logstash-config
  namespace: kube-logging
  labels:
    app: logstash
data:
  logstash.conf: |-
    input {
      tcp {
        port => 5000
        type => syslog
      }
    }

    filter {
        grok {
            match => {"message" => "%{SYSLOGLINE}"}
        }
    }

    output {
      elasticsearch {
        hosts => ["http://elasticsearch:9200"]
        index => "syslog-%{+YYYY.MM.dd}"
      }
      stdout { codec => rubydebug }
    }

---

kind: Deployment
apiVersion: apps/v1beta1
metadata:
  name: logstash
  namespace: kube-logging
  labels:
    app: logstash
spec:
  template:
    metadata:
      labels:
        app: logstash
    spec:
      containers:
      - name: logstash
        image: docker.elastic.co/logstash/logstash:7.2.1
        imagePullPolicy: Always
        env:
        - name: ELASTICSEARCH_HOST
          value: elasticsearch
        - name: ELASTICSEARCH_PORT
          value: "9200"
        - name: ELASTICSEARCH_USERNAME
          value: elastic
        - name: ELASTICSEARCH_PASSWORD
          value: changeme
        - name: ELASTIC_CLOUD_ID
          value:
        - name: ELASTIC_CLOUD_AUTH
          value:
        ports:
        - name: logstash
          containerPort: 5000
          protocol: TCP
        securityContext:
          runAsUser: 0
        resources:
          limits:
            memory: 800Mi
          requests:
            cpu: 100m
            memory: 100Mi
        volumeMounts:
        - name: config
          mountPath: /usr/share/logstash/pipeline/logstash.conf
          readOnly: true
          subPath: logstash.conf
      volumes:
      - name: config
        configMap:
          defaultMode: 0600
          name: logstash-config

---

kind: Service
apiVersion: v1
metadata:
  name: logstash
  namespace: kube-logging
  labels:
    app: logstash
spec:
  selector:
    app: logstash
  clusterIP: None
  ports:
    - name: tcp-port
      protocol: TCP
      port: 5000
      targetPort: 5000
---

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: logstash-external
  namespace: kube-logging
  labels:
    app: logstash
  annotations:
    kubernetes.io/ingress.class: traefik
    traefik.ingress.kubernetes.io/frontend-entry-points: tcp
spec:
  rules:
  - host: logstash.dev.domain.com
    http:
      paths:
      - backend:
          serviceName: logstash
          servicePort: 5000

able to send test message :

echo -n "test message" | nc logstash.dev.domain.com 5000

But don't see anything for tcpdump port 5000 in logstash container.

If I run echo -n "test message" | nc logstash.dev.domain.com 5000 from logstash container, then I see this message showing up for tcpdump port 5000 on logstash container.

Within EKS from any container I can send test message echo -n "test message 4" | nc -q 0 logstash 5000 and its received by logstash and pushed to ElasticSearch.

But not from outside of the cluster. So looks like traefik ingress controller is the issue here.

I have traefik ingress controller for EKS.

traefik.toml: |
  defaultEntryPoints = ["http","https"]
  logLevel = "INFO"
  [entryPoints]
    [entryPoints.http]
      address = ":80"
      compress = true
      [entryPoints.http.redirect]
      entryPoint = "https"
      [entryPoints.http.whiteList]
      sourceRange = ["0.0.0.0/0""]
    [entryPoints.https]
      address = ":443"
      compress = true
      [entryPoints.https.tls]
      [entryPoints.https.whiteList]
      sourceRange = ["0.0.0.0/0"]
    [entryPoints.tcp]
      address = ":5000"
      compress = true

and Service :

kind: Service
apiVersion: v1
metadata:
  name: ingress-external
  namespace: kube-system
  annotations:
    service.beta.kubernetes.io/aws-load-balancer-type: "nlb"
spec:
  type: LoadBalancer
  externalTrafficPolicy: Local
  selector:
    app: traefik-ingress-lb
  ports:
    - name: http
      protocol: TCP
      port: 80
      targetPort: 80
    - name: https
      protocol: TCP
      port: 443
      targetPort: 443
    - name: tcp-5000
      protocol: TCP
      port: 5000
      targetPort: 5000

What is wrong here ?

Can you verify that the logs from your raspberries reach logstash? Can you see the logs on the stdout? — apt-get_install_skill, Sep 17 '19 at 17:24

score 0 · Answer 1 · answered Sep 13 '19 at 11:51

If you haven't used logstash before in the past then you may need to manually create a logstash index. The data won't appear under filebeat as elasticsearch isn't receiving the data from filebeat but logstash itself. I may be wrong in this answer altogether. However, if you go to:

Settings > Index patterns > Create Index pattern Then proceed to type in logstash where it asks for a name and select logstash from underneath like so:

After creating this you should then get a drop down on the Discover page that says logstash. Under the logstash drop down you should see all of the data you are pushing through

You may already have a logstash index setup and this may not be the issue at all

There is no index matching `logstash-*`. I think it's because data is not getting pushed to ElasticSearch. — roy, Sep 13 '19 at 14:09

score 0 · Answer 2 · answered Sep 17 '19 at 17:08

0

The hosts setting in the elasticsearch output looks suspicious to me. As stated in the documentation one have to specify an URI there. You should specify the instance with the protocol like

http(s)://{IP or name}:9200

like you do in your curl request.

I would give that a try.

answered Sep 17 '19 at 17:08

apt-get_install_skill

2,818
10
27

Tried with `http://elasticsearch:9200`, still not messages in `logstash` – roy Sep 17 '19 at 17:21

traefik ingress not forwarding tcp messages

2 Answers2