Aerogear App Metrics

This is the server component of the AeroGear Metrics Service. It is a RESTful API that allows mobile clients to send metrics data which will get stored in a PostgreSQL database. The service is written in Golang.

Prerequisites

See the Contributing Guide for more information.

Getting Started

This section documents the ideal setup for local development. If you'd like to simply run the entire application in docker-compose, check out the relevant section.

Golang projects are kept in a workspace that follows a very specific architecture. Before cloning this repo, be sure you have a $GOPATH environment variable set up.

Clone the Repsitory

git clone git@github.com:aerogear/aerogear-app-metrics.git $GOPATH/src/github.com/aerogear/aerogear-app-metrics

Install Dependencies

make setup

Note this is using the dep package manager under the hood. You will see the dependencies installed in the vendor folder.

Start the Database

docker-compose up db

Start the Server

go run cmd/metrics-api/metrics-api.go

2018/02/27 10:51:54 Starting application... going to listen on :3000

The application's default configuration will allow it to connect to the database created by docker-compose.

You can test out the healthcheck endpoint:

curl -i http://localhost:3000/healthz

HTTP/1.1 200 OK
Content-Type: application/json; charset=UTF-8
Date: Tue, 27 Feb 2018 10:53:16 GMT
Content-Length: 56

{"time":"2018-02-27T10:53:16.313301415Z","status":"ok"}

Run Entire Application with Docker Compose

This section shows how to start the entire application with docker-compose. This is useful for doing some quick tests (using the SDKs) for example.

First, compile a Linux compatible binary:

make build_linux

This binary will be used to build the Docker image. Now start the entire application.

docker-compose up

Example Client Requests

This section shows example curl requests which can be used to send some data to the /metrics endpoint.

App and Device Related Metrics

The following request corresponds to an app initialization event.

Init type metric request

curl -i -X POST \
  http://localhost:3000/metrics \
  -H 'Cache-Control: no-cache' \
  -H 'Content-Type: application/json' \
  -H 'Postman-Token: 87bf2b99-7cdc-8df9-9b2d-6cdcd2932159' \
  -d '{
  "clientId": "453de7432",
  "type": "init",
  "data": {
    "app": {
      "appId": "com.example.someApp",
      "sdkVersion": "2.4.6",
      "appVersion": "256",
      "framework": "native"
    },
    "device": {
      "platform": "android",
      "platformVersion": "27"
    }
  }
}'

Security Related Metrics

The following request corresponds to on-device security check events.

curl -X POST \
  http://localhost:3000/metrics \
  -H 'Cache-Control: no-cache' \
  -H 'Content-Type: application/json' \
  -H 'Postman-Token: 6db8f0ae-4e9e-1ae1-530b-138b1a87bfe7' \
  -d '{
  "clientId": "9796fb69-a566-43fb-b932-03951d82437f",
  "type": "security",
  "timestamp": 1519989859342,
  "data": {
    "security": [
      {
        "id": "org.aerogear.mobile.security.checks.DeveloperModeCheck",
        "name": "Developer Mode Check",
        "passed": true
      },
     {
        "id": "org.aerogear.mobile.security.checks.EmulatorCheck",
        "name": "Emulator Check",
        "passed": false
      },
      {
        "id": "org.aerogear.mobile.security.checks.DebuggerCheck",
        "name": "Debugger Check",
        "passed": false
      },
      {
        "id": "org.aerogear.mobile.security.checks.RootedCheck",
        "name": "Rooted Check",
        "passed": false
      },
      {
        "id": "org.aerogear.mobile.security.checks.ScreenLockCheck",
        "name": "Screen Lock Check",
        "passed": false
      }
    ]
  }
}
'

Successful requests will return an empty 204 response.

If you have the psql command line tools you can connect to the Database and verify the data was inserted.

psql -U postgresql -d aerogear_mobile_metrics --host localhost
Password for user postgresql: # postgres

aerogear_mobile_metrics=> select * from mobileappmetrics;

Builds and Testing

The makefile provided provides commands for building and testing the code. For now, only the most important commands are documented.

make setup - Downloads the application dependencies.
make build - Compiles a binary for your current system. The binary is output at ./aerogear-app-metrics.
make build_linux - Compiles a Linux compatible binary. The binary is output at ./dist/linux_amd64/aerogear-app-metrics. This is mainly used for Docker builds. make build can still be used if you are a Linux user.
make docker_build - Builds a Binary from source and uses that binary to create a Docker image.
make test - Runs the unit tests (alias for make test-unit).
make test-integration - Runs unit tests and integration tests. This requires a running database server.
make test-integration-cover - Same as make test-integration but also generates a code coverage report. Used in the CI service.

Generating test data

Use the scripts/data-fill script to generate random data targetting the same database as the main binary. By default it generates 1k records with some data variance. You can override these via cli flags:

go run scripts/data-fill.go -n=100 -apps=5

It is also possible to specify a given number of records of each metrics type to generate using the n* flags:

go run scripts/data-fill.go -nInit=100 -nSecurity=1000

The above will create 100 entries of records containing the app and platform keys inside data, followed by 1000 containing security.

Environment Variables

The aerogear-app-metrics server is configured using environment variables:

Server Configuration

Variable	Default	Description
PORT	3000	The port the server will listen on
LOG_LEVEL	info	Can be one of [debug, info, warning, error, fatal, panic]
LOG_FORMAT	text	Can be one of [text, json]
DBMAX_CONNECTIONS	100	The maximum number of concurrent database connections the server will open

Database Connection Parameters

The database connection is configured using the table of environment variables below. These environment variables correspond to the PostgreSQL libpq environment variables. The table below shows all of the environment variables supported by the pq driver used in this server.

Variable	Default	Description
PGDATABASE	aerogear_mobile_metrics	The database to connect to
PGUSER	postgresql	The database user
PGPASSWORD	postgres	The database password
PGHOST	localhost	The database hostname to connect to
PGPORT	5432	The database port to connect to
PGSSLMODE	disable	The SSL mode
PGCONNECT_TIMEOUT	5	The default connection timeout (seconds)
PGAPPNAME		The application_name connection parameter
PGSSLCERT		The sslcert connection parameter.
PGSSLKEY		The sslkey connection parameter.
PGSSLROOTCERT		The sslrootcert connection parameter

How to Release a New Version

The release process for Aerogear maintainers is very simple. From the Github UI, simply create a release. The release tag and title must be in the format x.y.z. Formats such as 1.0 or v1.0.0 are not valid.

This will kick off an automated process in the CI service, where the following steps occur:

The release tag is checked out.
The code is built and the full test suite is run.
The goreleaser tool generates a Changelog and binaries for Windows, MacOS and Linux. These are added to the release created in Github.
A new docker image is built and given the tags latest and <git tag> (where <git tag> is the x.y.z tag that was used).
The docker image is pushed to the Aerogear organization in Github.

The automated release process takes 2-3 minutes to complete on average.

Using Swagger UI

A Swagger UI can be used for testing the aerogear-app-metrics service. The go-swagger tools must be installed:

go get -u github.com/go-swagger/go-swagger/cmd/swagger

Then generate the Swagger spec as follows:

make generate

This creates swagger.json file in the root of the project.

Run the Swagger UI using the official image. This mounts the root of the project as a volume.

docker run -p 8080:8080 -e SWAGGER_JSON=/etc/swagger/swagger.json -v `pwd`:/etc/swagger swaggerapi/swagger-ui

The Swagger UI is available at localhost:8080.

How it Works

Operations are annotated in router.go e.g. metrics, healthz, ping
- These annotations define what the endpoint is, its method, expected content type, and the request body (via a $ref usually)
Models are annotated Go structs in types.go. The swagger spec generator reads the structs to generate the spec.

Known Issues

The timestamp field is a json.Number. This gets generated as a string field in the swagger spec. (Technically this is not an issue because the json.Number type is actually a string.)
You may see caching issues in the browser when the swagger.json file changes (not sure why/when this happens). Clearing local storage in the browser (via developer tools) should fix it.

Contributing

All contributions are hugely appreciated. Please see our Contributing Guide for guidelines on how to open issues and pull requests. Please check out our Code of Conduct too.

Any Questions?

There are a number of ways you can get in in touch with us:

Open a Github issue.
Start a thread in the Aerogear Mailing List (Open to anyone).
Reach out to us on IRC. The Aerogear team can be found at the #Aerogear channel on freenode.net (Open to anyone).

aerogear-attic/aerogear-app-metrics