cs

Container Solution API-exercise

Container Solution API-exercise, to assess technical proficiency with Software Engineering, DevOps, and Infrastructure tasks.

The Titanic API

The Titanic API is written in golang. It leverages go-kit to grant better modularity and micro-services support out-of-the-box.

The data layer is designed around CockroachDB, deployed to the K8S cluster leveraging a StatefulSet configuration.

Build the API

There are two ways here available to build the API code; a targetted method and a cross-plattform builder script; both allow to create portable executables, compatible with Alpine Linux, compiled statically linking C bindings -installsuffix 'static', and omitting the symbol and debug info -ldflags "-s -w".

Targetted build, based on your system/architecture

# change according to your system/architecture
CGO_ENABLED=0 GOARCH=[amd64|386] GOOS=[linux|darwin] go build -ldflags="-w -s" -a -installsuffix 'static' -o cs cmd/cs/main.go

Cross-platform build, leveraging the build.sh script

chmod +x build.sh && ./build.sh

The build.sh script will also re-build and push the docker images to our private GCR.

Currently, the builds in the releases folder are available for the following platforms and architectures:

• darwin / amd64;
• darwin / 386;
• linux / amd64;
• linux / 386.

Build the Docker images locally

The Docker images are in the docker folder. To build the production API Docker image, run these commands in your terminal:

cd docker
docker build --no-cache --file Dockerfile -t gcr.io/$PROJECT_ID/cs-api:latest .

To build the devlopment Docker image you must extend the build context to the top-level folder of this repo, and include the files go.mod and go.sum, along with the source code and run the API:

# the Dockerfile is outside the build context
docker build --no-cache --file ./docker/dev.Dockerfile -t gcr.io/$PROJECT_ID/cs-api:dev .

Run the API locally in Docker

Running the API locally is quite simple; it does not require any particular language or library installed on your system, other than Docker. The golang server is listening both on port 8443/TCP over TLS and 3000/TCP over http; to properly run the API locally, before spawning the docker image, you need to generate the certs as follow, from the top dir of this repo:

mkdir -p $(pwd)/tls; \
openssl req -new -newkey rsa:4096 -days 365 -nodes -x509 \
-subj "/C=NL/ST=Amsterdam/L=Amsterdam/O=hyperd/CN=cs-api.hyperd.sh" \
-keyout $(pwd)/tls/tls.key -out $(pwd)/tls/tls.crt; \
chmod 444 tls/*

Once the cert and the key are in place, it's enough to lift the docker image with the command described here below, mounting the correct volume, to have the TLS configured correctly:

# production | develpment
docker run -d --name cs-api -v $(pwd)/tls:/etc/tls/certs -p 3000:3000 -p 8443:8443  gcr.io/$PROJECT_ID/cs-api:[latest|dev]

API Walkthrough

The Titanic API exposes the following methods:

POST /people/ adds another passenger to the people collection:

payload='
{
  "survived": true,
  "pclass": 1,
  "name": "Francesco",
  "sex": "M",
  "age": 30,
  "siblings_spouses_abroad": false,
  "parents_children_aboard": false,
  "fare": 7.34
}
'
curl -d "$payload" -H "Content-Type: application/json" -X POST http://localhost:3000/people/
# {}

GET /people/:uuid retrieves the given passenger by uuid from the people collection:

curl http://localhost:3000/people/35d4ab59-fa9d-478d-a57e-61b526ee0a33 | jq
{
  "people": {
    "uuid": "35d4ab59-fa9d-478d-a57e-61b526ee0a33",
    "survived": true,
    "pclass": 1,
    "name": "Francesco",
    "sex": "M",
    "age": 30,
    "siblings_spouses_abroad": false,
    "parents_children_aboard": false,
    "fare": 7.34
  }
}

DELETE /people/:uuid removes the given passenger:

curl -X "DELETE" http://localhost:3000/people/35d4ab59-fa9d-478d-a57e-61b526ee0a33
# {}

PATCH /people/:uuid partial update of the passenger information:

payload='
{
  "siblings_spouses_abroad": true,
  "parents_children_aboard": true
}
'
curl -d "$payload" -H "Content-Type: application/json" -X PATCH -k http://localhost/people/35d4ab59-fa9d-478d-a57e-61b526ee0a33

PUT /people/:uuid posts updated information about a given passenger:

payload='
{
  "survived": true,
  "pclass": 1,
  "name": "Francesco",
  "sex": "M",
  "age": 30,
  "siblings_spouses_abroad": true,
  "parents_children_aboard": true,
  "fare": 9.81
}
'
curl -d "$payload" -H "Content-Type: application/json" -X PUT http://localhost:3000/people/35d4ab59-fa9d-478d-a57e-61b526ee0a33

GET /people/ retrieves all the passengers of the Titanic:

curl http://localhost:3000/people/ | jq
{
  "people": [
    {
      "uuid": "30615024-ada8-4af6-8611-882c006d17f4",
      "survived": true,
      "pclass": 1,
      "name": "Francesco",
      "sex": "M",
      "age": 30,
      "siblings_spouses_abroad": false,
      "parents_children_aboard": false,
      "fare": 7.34
    },
    {
      "uuid": "363f558a-eeb1-4bf6-b570-33e61e60b867",
      "survived": true,
      "pclass": 1,
      "name": "Anne McLeod",
      "sex": "F",
      "age": 49,
      "siblings_spouses_abroad": true,
      "parents_children_aboard": true,
      "fare": 9.34
    },
    ...
  ]
}

Deploy the API to GCP

To deploy the stack to GKE on GCP follow this documentation.