opencensus-php-docker

OpenCensus PHP Ecosystem using Docker Compose for quick prototyping & testing.

This docker-compose demo uses the following container set-up:

                                                                                       ┌────────┐
┌────────────┐                ┌───────────┐         ┌────────────┐          ┌─────────▶│ Zipkin │
│ PHP APACHE │──unix socket──▶│ OC Daemon │──gRPC──▶│  OC Agent  │──export──┤          └────────┘
└────────────┘                └───────────┘         └────────────┘          │          ┌────────────┐
                                                                            └─────────▶│ Prometheus │
                                                                                       └────────────┘

Starting the ecosystem

To start the ecosystem, simply run:

docker-compose up

After a while you should have the following available services:

service	description	URL
Zipkin	Tracing UI	http://localhost:9411
Prometheus	Metrics UI	http://localhost:9090
OC Agent	zPages	http://localhost:55679/debug/tracez http://localhost:55679/debug/rpcz
OC Agent	Prometheus scrape address	http://localhost:8888/metrics
Apache PHP	Apache PHP Server	http://localhost:8080

If one of the above ports is already in use, you can adjust the ports in the docker-compose.yml file e.g.

  ports:
    # Alternate host port 9190 for the Prometheus UI
    - 9190:9090

Running example code

The data/php folder is the PHP container's document root and the data/php/html directory is the Apache web root. This allows you to have code live outside the Apache path (e.g. a compose vendor directory).

A small example script has been added as data/php/html/helloworld/index.php. To run it you need to install the code's dependencies first with composer.

The Apache PHP container which by default is named ocphp_php has composer installed so getting the dependencies ready can be done by running:

docker exec -ti ocphp_php composer install

After this a vendor directory will be available under data/php and contain the OpenCensus library and its dependencies.

The script is similar to the typical Helloworld examples found in other OpenCensus language libraries to highlight how similar usage is in a polyglot environment while still trying to be idiomatic for each different language.

Using Stackdriver

If you wish to use Stackdriver as your tracing and/or metrics backend you need to adjust the agent configuration file config.yaml. Uncomment the lines showing the stackdriver set-up and adjust to the GCP project id you wish to use:

zipkin:
    endpoint: "http://zipkin:9411/api/v2/spans"
    # Uncomment below and provide correct projectid to enable stackdriver export.
    stackdriver:
        project: "my-project-id"
        enable_tracing: true

To be able to connect to Stackdriver you need a credentials file. See: https://cloud.google.com/docs/authentication/getting-started for more information.

Once you have a credentials file, place it in the data/agent/credentials directory. Finally adjust the docker-compose.yml and modify value for the environment variable GOOGLE_APPLICATION_CREDENTIALS as found in the agent container configuration. If your file is called super-secret.json you'd want something like this:

    agent:
        image: basvanbeek/opencensus-agent:latest
        container_name: ocphp_agent
        environment:
            # Set to the correct credentials file if using Stackdriver
            GOOGLE_APPLICATION_CREDENTIALS: credentials/super-secret.json

Now you can simply run:

docker-compose up

And your data is exported to Stackdriver in your GCP project.