Buoyant Books App

This is a sample distributed (microservices) Ruby app using Sinatra, ActiveRecord, and ActiveResource. The app is designed to demonstrate the various value propositions of Linkerd 2.0 including debugging, observability, and monitoring. Some of the services in the app periodically fail. This is by design in order to demo debugging and monitoring in Linkerd 2.0.

The application is composed of the following four services:

• webapp.rb
• authors.rb
• books.rb
• traffic/main.go (demo traffic generator, written in Go)

---

Running in Kubernetes

You can deploy the application to Kubernetes using the Linkerd 2.0 service mesh.

1. Install the linkerd CLI

   curl https://run.linkerd.io/install | sh

2. Install the Linkerd control plane

   linkerd install | kubectl apply -f -

3. Inject and deploy the application

   curl https://run.linkerd.io/booksapp.yml | linkerd inject - | kubectl apply -f -

4. Use the app!

   kubectl port-forward svc/webapp 7000
   open "http://localhost:7000"

5. View the Linkerd dashboard!

   linkerd dashboard

Running with MySQL

The default booksapp configuration uses SQLite. It's also possible to run the app with a MySQL backend, using the configs in the k8s/ directory. The MySQL configuration uses a separate pod for the storage backend, which allows running multiple replicas of each of the app deployments.

1. Start by installing the MySQL backend

   kubectl apply -f k8s/mysql-backend.yml

2. Verify that the mysql-init job successfully completes

   kubectl get po
   NAME                    READY     STATUS      RESTARTS   AGE
   mysql-9bd5bcfdf-7jb2s   1/1       Running     0          3m
   mysql-init-29nxv        0/1       Completed   0          3m

3. Install Linkerd as described above; install the app configured to use MySQL

   linkerd install | kubectl apply -f -
   linkerd inject k8s/mysql-app.yml | kubectl apply -f -

4. Use the app!

   kubectl port-forward svc/webapp 7000
   open "http://localhost:7000"

Deploy Chaos Monkey

This repo includes a Chaos Monkey script, that randomly kills a pod every 10 seconds. It is intended to be run with the Kubernetes / MySQL configuration. To deploy, run:

```bash
kubectl apply -f k8s/mysql-chaos.yml
```

---

Service Profiles

In order to record per-route metrics, you can create service profiles for the webapp, books, and authors services based on their Swagger specs:

```bash
linkerd profile --open-api swagger/webapp.swagger webapp | kubectl apply -f -
linkerd profile --open-api swagger/books.swagger books | kubectl apply -f -
linkerd profile --open-api swagger/authors.swagger authors | kubectl apply -f -
```

You can then view route data for each service:

```bash
linkerd routes webapp
```

```bash
linkerd routes books
```

```bash
linkerd routes authors
```

---

Traffic Splits

You can use a modified version of booksapp to demo a trafficsplit.

1. Install booksapp-trafficsplit.yml which includes the MySQL backend, a modified version of booksapp.yml that does not introduce a FAILURE_RATE, a second authors service called authors-clone, and two trafficsplits. Linkerd has already been injected into booksapp and authors-clone.

   kubectl apply -f booksapp-trafficsplit.yml

2. Verify that two trafficsplits now exist in the default namespace.

   kubectl get ts

3. Give the app 1-2 minutes to begin sending traffic.

4. Run watch linkerd stat ts to verify that the trafficsplits are working. You should see the below results -- one trafficsplit dividing traffic 50/50 between authors and authors-clone, one dividing traffic 100/0 between webapp and the non-existent webapp-clone (You can create a split between two services before the second service has been created).

   NAME            APEX      LEAF            WEIGHT   SUCCESS      RPS   LATENCY_P50   LATENCY_P95   LATENCY_P99
   authors-split   authors   authors           500m   100.00%   3.1rps           8ms          29ms          37ms
   authors-split   authors   authors-clone     500m   100.00%   3.5rps           8ms          23ms          37ms
   webapp-split    webapp    webapp               1   100.00%   7.3rps          26ms          74ms          95ms
   webapp-split    webapp    webapp-clone         0         -        -             -             -             -
   

5. When you are done, delete the app.

   kubectl delete -f booksapp-trafficsplit.yml

Running Locally

You can also run the application locally for development.

1. Create, migrate, and seed the database

   bundle install
   bundle exec rake db:create
   bundle exec rake db:migrate
   bundle exec rake db:seed

2. Start the web app

   bundle exec rake dev:webapp

3. Start the authors app

   bundle exec rake dev:authors

4. Start the books app

   bundle exec rake dev:books

5. Open the website

   open "http://localhost:7000"

Administration

Docker

All of the Docker images used for this application are already published publicly and don't need to be built by hand. If you'd like to build the images locally follow the instructions below.

1. Build the buoyantio/booksapp image

   docker build -t buoyantio/booksapp:latest .

2. Build the buoyantio/booksapp-traffic image

   docker build -t buoyantio/booksapp-traffic:latest traffic