Skip to content

Latest commit

 

History

History
270 lines (171 loc) · 6.89 KB

README.md

File metadata and controls

270 lines (171 loc) · 6.89 KB
Raw

elasticsearch full monitoring platform

This repo contains the source code for deploying full elasticstack + running application workload on a kubernetes cluster. You can do it step by step using provided helm charts, or deploy them all running the pulumi script. Yes, there is one line command too:

docker run -v ${HOME}/.kube:/usr/multicloud/kube iiglesias/multicloud-sandbox:0.1.0

If you have a kubernetes cluster try it out

images/

images/

images/

Execute

Pulumi

Pulumi is a modern IaC based on Hashicorp´s Terraform.

Its mainly used for cloud provisioning but in this case we will be using just the kubernetes provider for deploying our stack on a given kube/config

You can access helm templates https://github.com/iiglesiasg/helm-charts-repo

To run the scrip you will need following tools:

By default pulumi run the stack against your local .kube/config. Your config should be pointing the kubernetes cluster

You should login on pulumi and create your account. Then run:

npm install ## for downloading pulumi´s libraries
pulumi login
pulumi up

An stack name may be required, type a name if prompted

It will preview the stack to be deployed and prompt for confirmation

You can delete the stack by running

pulumi destroy

Docker

I made a docker image so you dont need to install so much tools.

Only needed is a kube/config pointing to the desired kubernetes cluster

Simply run:

docker run -v ${HOME}/.kube:/usr/multicloud/kube iiglesias/multicloud-sandbox:0.1.0
  • Note that pulumi destroy is not implemented

The dockerfile is included in this repo. For pulumi´s stack status management I recommend creating your own dockerfile with your pulumi token

WHAT IS DEPLOYED

ECK

  • elasticsearch database cluster
  • elastic APM collector
  • kibana

Elastic tools

  • logstash
  • MetricBeats
    • k8s
    • Infra
    • Mongo
    • elastic cluster health

Application Workload

This workload is composed by 3 services

images/

SpringBoot services:

curl --header "Content-Type: application/json" --request POST http://{ADAPTER_IP}}:8090/MONADTST/testhttp -d @body.json

Where the body

{
    "adapter":{
            "msg": "adapter",
            "statusCode": 200,
            "delayMillis": 1,
            "encapsulateError" : false
    },
    "composite":{
            "msg": "composite",
            "statusCode": 200,
            "delayMillis": 200, 
            "encapsulateError" : false
    },
    "core": {
            "msg": "core",
            "statusCode": 200,
            "delayMillis": 3,
            "encapsulateError" : false
    }
}

delayMillis: For elastic APM inferred span

Mongodb

Core service will perform insert and querys on a mongodb

INGRESS-CONTROLLER + CERT-MANAGER

Just a minimal configuration for exposing the workload throught TLS by creating ingress manifest.

It is intended to serve ACME certificates once an fqdn is assigned to the Nginx. By default clusterIssuers will be created but only selfsigned certificates are issued.

sudo vi /etc/hosts
# Add load balancer´s ip for the ingress host spec
104.155.22.7	adapter.multicloud.inditex.com

curl -kv https://adapter.multicloud.inditex.com/MONADTST/amiga/healthcheck

apiVersion: networking.k8s.io/v1beta1
kind: Ingress
metadata:
  name: adapter-helm
  annotations:
    kubernetes.io/ingress.class: nginx
    cert-manager.io/cluster-issuer: selfsigning-issuer ## For a different clusterIssuer 
spec:
  tls:
  - hosts:
    - adapter.multicloud.inditex.com
    secretName: tls-secret
  rules:
  #- host: adapter.multicloud.inditex.com
  - http:
      paths:
      - backend:
          serviceName: adapter-helm
          servicePort: 8090
        path: /MONADTST

Load Generator

Once the cluster is created a Kubernetes Job will perform load request to our demo workload.

This load Generator is based on Vegeta and will be generating load (distributed traces) for the first 15 mins.

Base image was provided by https://github.com/peter-evans/vegeta-docker

KIBANA

Not everything is automated, some instructions must be followed to make the full platform running

Login

kubectl get secret quickstart-es-elastic-user -o yaml
## Decode the base64 value of elastic user
echo "OE1xNTJwVmkycEJvUlk0MXh5MjdlZDk0" | base64 -d
kubectl port-forward svc/quickstart-kb-http 5601:5601

Type https://localhost:5601 on your browser

images/

  • User: elastic
  • password: ${DECODED_BASE64}

Check elasticsearch index and reload them

images/

Create kibana index

  • logstash-*
    • Application workload logs
  • metrics-*
    • Our application workload is using this index for Micrometer

[^]: MetricBeat and apm index are automatically created

images/

images/

Stack Monitoring

images/

Turn on monitoring as metricBeat is already configured

images/

LOGS

In order to use kibana logs, we must set logs index to our newly created kibana index logstash-*

Change source configuration

images/

Replace filebeat by logstash index

images/

Now you can correlate traces and logs

images/

ENJOY

  • Add Nginx module to MetricBeat

    • MetricBeat configuration is stored in a configmap

  • Try kibana lens new dashboard engine

  • Update elasticsearch node count and storage size

    • update elasticsearch

      kubectl edit elasticsearch

    • View changes on Kibana´s Stack Monitoring