Skip to content

Latest commit

 

History

History
448 lines (301 loc) · 12.7 KB

00-setup-introduction.md

File metadata and controls

448 lines (301 loc) · 12.7 KB
Raw

Introducing the setup

Learning goal

  • Try out the sentences application
  • Add the applications traffic to the istio service mesh
  • Familiarize yourself with the Kiali management console

Introduction

This exercise introduces the sentences application which you will be using during the course. It also introduces you to the Kiali management console for the Istio service mesh.

In the beginning of the exercise you will deploy the sentences application and generate traffic between the services. But there will be no Envoy sidecars injected and Kiali will not be able to observe the traffic.

Afterwards you will use a couple of different methods to enable sidecars. This will allow Kiali to observe the traffic.

💡 More Info on the micro service application and Kiali

Sentences application

This application implements a simple 'sentences' builder, which can build sentences from the following simple algorithm:

age = random(0,100)
name = random(['Peter','Ray','Egon'])
return name + ' is ' + age + ' years'

The application is made up of three services, one which can be queried for the random age, one which can be queried for a random name and a frontend sentence service, which calls the two other through HTTP requests and formats the final sentences.

The source code for the application can be seen in the sentences-app/ folder.

Kiali

Kiali provides dashboards and observability by showing you the structure and health of your service mesh. It provides detailed metrics, Grafana access and integrates with Jaeger for distributed tracing.

One of it's most powerful features are it's graphs. They provide a powerful way to visualize the topology of your service mesh.

It provides four main graph renderings of the mesh telemetry.

  • The workload graph provides a detailed view of communication between workloads.

  • The app graph aggregates the workloads with the same app labeling, which provides a more logical view.

  • The versioned app graph aggregates by app, but breaks out the different versions providing traffic breakdowns that are version-specific.

  • The service graph provides a high-level view, which aggregates all traffic for defined services.

Kiali overview

We are using Kiali to visualize the work done in this Istio course.

Exercise: Get Familiar with Istio and Kiali.

In this exercise, we are first deploying our application with vanilla Kubernetes. We then visit the Kiali website to see that without a sidecar our service will not be included in istio.

Then you will be using a few different techniques to enable Istio sidecars and see the traffic flowing in Kiali.

Overview

A general overview of what you will be doing in the Step By Step section.

  • Deploy the sentences application services with kubectl

  • Producing traffic to the sentences application

  • Investigating Kiali

  • Enabling sidecars with automatic sidecar injection

  • Enabling sidecars with annotations

  • Manually force injection of sidecars

Step By Step

Expand the Tasks section below to do the exercise.

Tasks

Task: Inspect your namespace

You will need to know your namespace for later exercises. It is provided for you in the environment variable STUDENT_NS. Check it with the following command.

echo $STUDENT_NS

Execute the following command and make sure it matches the value of the environment variable STUDENT_NS.

kubectl config view --output 'jsonpath={..namespace}'; echo

Task: Deploy the sentences application

Open a terminal in the root of the git repository (istio-katas) and use kubectl to deploy v1 of the application.

kubectl apply -f 00-setup-introduction/

Task: Observe the number of services and pods running

kubectl get pod,svc

Once all the pods are running you should see something like. It may take a few seconds.

NAME                             READY   STATUS    RESTARTS   AGE
pod/age-7976688957-mbvzz         1/1     Running   0          2s
pod/name-v1-587b56cdf4-rwcwt     1/1     Running   0          2s
pod/sentences-6dffccb8c6-7fd57   1/1     Running   0          2s

NAME                TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
service/age         ClusterIP   172.20.123.133   <none>        5000/TCP         2s
service/name        ClusterIP   172.20.108.51    <none>        5000/TCP         2s
service/sentences   NodePort    172.20.168.218   <none>        5000:30326/TCP   2s

Task: Run the loop-query.sh script

Run the following to continuously query the sentence service and observe the output.

Do this is a new terminal.

./scripts/loop-query.sh

You should see output from the shell that looks like this.

Using 10.0.36.130:30459, header ''
Egon is 40 years.
Ray is 16 years.
Peter is 38 years.
Peter is 19 years.
Ray is 66 years.

Traffic is now flowing between the services. But that doesn't mean it is part of the service mesh yet...

Task: Open Kiali and find the sentences application

Browse to Applications on the left hand menu. Click Namespace drop-down at the top left and enter your namespace.

Finally, select your sentences application from the center-part of the UI.

You will see the application, workloads and services are discovered by Kiali. But not much else.

The red icons beside the workloads mean we have no istio sidecars deployed. Browse the different tabs to see that there is no traffic nor metrics being captured. As there are no sidecars the traffic is not part of the istio service mesh.

Sentences with no sidecars

Task: Pull sentences application down

kubectl delete -f 00-setup-introduction/

Task: Enable automatic sidecar injection

kubectl label namespace $STUDENT_NS istio-injection=enabled

Task: Redeploy sentences application

kubectl apply -f 00-setup-introduction/

Observe the number of services and pods running.

kubectl get pod,svc

You should see two containers per POD in ready state.

NAME                                READY   STATUS    RESTARTS   AGE
pod/age-v1-6fccc84ff-kkdgn          2/2     Running   0          4m4s
pod/name-v1-6644f45d6f-lndkm        2/2     Running   0          4m4s
pod/sentences-v1-5bbf7bcfcb-fphpp   2/2     Running   0          4m4s

NAME                TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
service/age         ClusterIP   172.20.228.238   <none>        5000/TCP         4m5s
service/name        ClusterIP   172.20.213.23    <none>        5000/TCP         4m4s
service/sentences   NodePort    172.20.106.197   <none>        5000:32092/TCP   4m4s

Run the following command to observe that an envoy proxy container has been injected into the application.

kubectl get pods -o=custom-columns=NAME:.metadata.name,CONTAINERS:.spec.containers[*].name

This should show an istio proxy sidecar for each service.

NAME                            CONTAINERS
age-v1-676bf56bdd-m6bcj         age,istio-proxy
name-v1-587b56cdf4-6tnhs        name,istio-proxy
sentences-v1-6ccc9fdcc5-fzt2g   sentences,istio-proxy

Task: Run the loop-query.sh script

./scripts/loop-query.sh

Task: Browse kiali and investigate the traffic flow

Browse to Applications on the left hand menu and select sentences.

Remember to filter by your namespace.

Now you can see there are sidecars and the traffic is part of the mesh.

  • Browse the different tabs to see the traffic and metrics being captured.

💡 It may take a minute before Kiali starts showing the traffic and metrics. You can change the refresh rate in the top right hand corner.

Sentences with sidecars

Task: Disable automatic sidecar injection for the age service

Edit the file 00-setup-introduction/age.yaml and add the annotation sidecar.istio.io/inject: 'false'.

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: sentences
    mode: age
    version: v1
  name: age-v1
spec:
  replicas: 1
  selector:
    matchLabels:
      app: sentences
      mode: age
      version: v1
  template:
    metadata:
      labels:
        app: sentences
        mode: age
        version: v1
      annotations:                          # Annotations block
        sidecar.istio.io/inject: 'false'    # True to enable or false to disable
    spec:
      containers:
      - image: praqma/istio-sentences:v1
        name: age
        ports:
          - containerPort: 5000
        env:
        - name: "SENTENCE_MODE"
          value: "age"

Apply the changes to the age.yaml file.

kubectl apply -f 00-setup-introduction/age.yaml

Use kubectl to see the number of pods running.

kubectl get pods

You should, eventually, see that the age service has only one pod. E.g. it no longer has a sidecar and is not part of the service mesh.

NAME                            READY   STATUS    RESTARTS   AGE
age-v1-574bfbb6b4-qb6rv         1/1     Running   0          3m50s
name-v1-795cf79f69-clrw4        2/2     Running   0          8m41s
sentences-v1-7cfbb658b6-rthxn   2/2     Running   0          8m41s

If you re-inspect the application graph in Kiali, you will also notice, that the age service is no longer being shown.

Automatic sidecar injection provides a pervasive and homogenous approach to ensuring the features istio provides. For example telemetry like metrics and traces for observability. If you do not want a sidecar for a service, use an opt out approach.

Task: Inject sidecar for the age service

To update the age deployment to again include a sidecar, w can use the following command:

cat 00-setup-introduction/age.yaml |grep -v inject | kubectl apply -f -

Use kubectl to see the number of pods running.

kubectl get pods

You should now see that the age service again has two pods. E.g. it has a sidecar and is again part of the service mesh.

NAME                            READY   STATUS    RESTARTS   AGE
age-v1-7b9f67b7dc-qxlxx         2/2     Running   0          76s
name-v1-795cf79f69-clrw4        2/2     Running   0          16m
sentences-v1-7cfbb658b6-rthxn   2/2     Running   0          16m

If you inspect the application graph in Kiali, you will see that the age service again is being shown.

You didn't modify the static yaml with the above command. You simply took the output of the cat command, piped it into grep which stripped out the annotation and applied the output with kubectl.

Task: Investigate the different graphs

Browse to the graphs and investigate the service, workload, app and versioned app graphs from the drop down at the top.

💡 Use the display options to modify what is shown in the different graphs. Showing request distribution is something we will be using often. Also ensure you are running the loop-query.sh script to generate traffic.

💡 Use the Legend button to explain the different objects being shown.

Graph Details

Summary

In this exercise you injected sidecars with automatic sidecar injection, disabled sidecars with an annotation and manually injected a sidecar from the command line. Manually injecting sidecars or using annotations is not a cohesive approach.

You were also introduced to the sentences application and Kiali. There is not enough time in the course to go into much more details around Kiali. But it has more features like the Istio Wizards feature which lets you create and delete istio configuration on the fly. It can do validation on the most common Istio objects and more.

See the documentation for a more complete overview.

Main takeaways are:

  • Annotations can be used to control sidecar injection.

  • Automatic sidecar injections is recommended. Automatic sidecar injection ensures a more pervasive and homogenous approach for traffic management and observability. It is less intrusive as it happens at the pod level and you won't see any changes to the yaml itself.

You can find more information about the different methods here.

And you can find more details about sidecar configuration here.

Cleanup

kubectl delete -f 00-setup-introduction/