Elasticsearch as a time-series storage

Built with PHP and Lumen (Laravel) framework

You can easily track any events type, single (f.e. mouse click, audio & video start, ad start & stop), repeatable (f.e. audio & video progress) and get reports for particular events and resources

What's included?

1. Events tracking
2. Reports:

• Total number of events by a requested event type for a given period
• Histogram for a given period by the given interval

Instalation

You can run app on an existing LAMP/LEMP stack or with the provided Docker compose configuration

Running inside Docker

1. Build containers $ cd /project_root && docker-compose build
2. Run docker $ docker-compose up -d
3. Login to the workspace container $ docker-compose run workspace bash Repeat steps from

Configure application

1. Create a new Elasticsearch index, from the project root: $ curl -XPUT "https://elastichost/indexname/typename" -d @database/esmappings.json
2. Install project dependencies $ php composer.phar install
3. Create and fill with the appropriate values your .env file $ cp .env.example .env
4. Cache routes $ php artisan route:cache

API Documentation

API documentation available with a http://swagger.io/

Usage

Events tracking (POST request parameters must be a valid JSON string)

$ curl -H "Content-Type: application/json" -X POST "https://example.com/api/stats/pulse" -d '{"event":"audiostart", "domain":"https://video.example.com", "uid":"123"}'

Response has a "session" key which can be used as a session for sending an associated events, f.e. audio progress for the same audio or video

Reporting API

Available at

• https://example.com/api/count - Total events
• https://example.com/api/histogram - Date histogram