MicroDonuts: An OpenTracing Walkthrough

Welcome to MicroDonuts! This is a sample application and OpenTracing walkthrough, written in Java.

OpenTracing is a vendor-neutral, open standard for distributed tracing. To learn more, check out opentracing.io, and try the walkthrough below!

Note that there are two git branches of note here.

• git checkout master illustrates a trivial multi-service app with cross-service tracing via OpenTracing
• git checkout no-tracing removes the tracing instrumentation, allowing the reader to add it in themselves

Step 0: Setup MicroDonuts

Getting it

Clone this repository and build the jar file (for this, Maven must be installed):

git clone [email protected]:opentracing-contrib/java-opentracing-walkthrough.git
cd java-opentracing-walkthrough/microdonuts
mvn package

Docker image

From the main directory:

cd java-opentracing-walkthrough/microdonuts
mvn package
cd ..
docker build . -t microdonuts

Running

Locally

MicroDonuts has two server components, API and Kitchen, which communicate each other over HTTP - they are, however, part of the same process:

cd java-opentracing-walkthrough/microdonuts
mvn package exec:exec

Docker

docker run -p 10001:10001 microdonuts

Accessing

In your web browser, navigate to http://127.0.0.1:10001 and order yourself some µ-donuts.

Pick a Tracer

Several OpenTracing-compatible Tracer implementations are supported out-of-the-box for convenience. Others can be added easily with a local change to App.java.

Jaeger

To run Jaeger locally (via Docker):

$ docker run -d -p 5775:5775/udp -p 16686:16686 jaegertracing/all-in-one:latest

Then add the following to microdonuts/tracer_config.properties:

tracer=jaeger
jaeger.reporter_host=localhost
jaeger.reporter_port=5775

Note that the all-in-one docker image presents the Jaeger UI at localhost:16686.

Zipkin

To run Zipkin locally (via Docker):

$ docker run -d -p 9411:9411 openzipkin/zipkin

Then add the following to microdonuts/tracer_config.properties:

tracer=zipkin
zipkin.reporter_host=localhost
zipkin.reporter_port=9411

Note that the all-in-one docker image presents the Zipkin UI at localhost:9411.

LightStep

If you have access to LightStep, you will need your access token. Add the following to microdonuts/tracer_config.properties:

tracer=lightstep
lightstep.collector_host=collector.lightstep.com
lightstep.collector_port=443
lightstep.access_token=XXXXXXXXXXXXXXX  // TODO: replace with your token

Step 1: Check out the no-tracing branch

The master branch in this repository has tracing instrumentation added as described below. To maximize your learnings, do a ...

git checkout no-tracing

... to start with a version of the code that's not instrumented yet. The guide below will let you learn-by-doing as you re-introduce that tracing instrumentation.

Step 2: Turnkey Tracing

When you go to add tracing to a system, the best place to start is by installing OpenTracing plugins for the OSS components you are using. Instrumenting your networking libraries, web frameworks, and service clients quickly gives you a lot of information about your distributed system, without requiring you to change a lot of code.

To do this, let's change the startup of the application to include tracing.

Start the GlobalTracer

In OpenTracing, there is a concept of a global tracer for everyone to access, As a convenience, we already have a function configureGlobalTracer that works with the MicroDonuts configuration file. In the main of microdonuts/src/main/java/com/otsample/api/App.java, right after the configuration file was loaded, we add:

        Properties config = loadConfig(args);
        if (!configureGlobalTracer(config, "MicroDonuts"))
            throw new Exception("Could not configure the global tracer");

After this, the tracer will be available globally through io.opentracing.GlobalTracer.get().

Instrument the outgoing HTTP requests

Our api component communicates with the kitchen one over HTTP using the OkHttp library, so we will instrument those requests using a middleware. In microdonuts/src/main/java/com/otsample/api/KitchenConsumer.java, inside the the KitchenConsumer constructor:

    TracingInterceptor tracingInterceptor = new TracingInterceptor(
             GlobalTracer.get(),
             Arrays.asList(SpanDecorator.STANDARD_TAGS));
    client = new OkHttpClient.Builder()
            .addInterceptor(tracingInterceptor)
            .addNetworkInterceptor(tracingInterceptor)
            .build();

Instrument the inbound HTTP kitchen server

Similarly, we will use a middleware to trace the incoming HTTP requests for the kitchen component. In microdonuts/src/main/java/com/otsample/api/KitchenContextHandler.java, inside the constructor do:

    TracingFilter tracingFilter = new TracingFilter(GlobalTracer.get());
    addFilter(new FilterHolder(tracingFilter), "/*", EnumSet.allOf(DispatcherType.class));

After this, the incoming requests to kitchen will be traced, too.

Check it out in your Tracer

Now that we're all hooked up, try ordering some donuts in the browser. You should see the traces appear in your tracer.

Search for traces starting belonging to the MicroDonuts component to see the patterns of requests that occur when you click the order button.

Step 3: Enhance

Now that the components in our system are linked up at the networking level, we can start adding application level tracing by tying multiple network calls together into a single trace.

In MicroDonuts, we'd like to know the time and resources involved with buying a donut, from the moment it is ordered to when it is delivered. Let's add OpenTracing's context to the requests from the api component to the kitchen one. In microdonuts/src/main/java/com/otsample/api/ApiContextHandler.java in OrderServlet do:

        @Override
        public void doPost(HttpServletRequest request, HttpServletResponse response)
            throws ServletException, IOException
        {
            Span orderSpan = GlobalTracer.get().buildSpan("order_span").start();
            request.setAttribute("span", orderSpan);

And then, at the end of this same method:

            Utils.writeJSON(response, statusRes);
            orderSpan.finish();
        }

Here we are creating a top level span that will be the parent of all the traced operations happening when ordering a set of donuts, and after that we store it in our HttpServletRequest object, so we can retrive this information next. In microdonuts/src/main/java/com/otsample/api/KitchenConsumer.java inside addDonut add a TagWrapper instance with the parent span:

    Span parentSpan = (Span) request.getAttribute("span");
    Request req = new Request.Builder()
        .url("http://127.0.0.1:10001/kitchen/add_donut")
        .post(body)
        .tag(new TagWrapper(parentSpan.context())) 
        .build();

This way, we are marking these requests as children of our main Span, and they will appear in the tracer properly organized, as belonging to the same operation.

And we're done! Buy some donuts, check out the spans under the MicroDonuts component and notice how the order and polling requests are now grouped under a single span, with timing information for the entire operation.

Step 4: Have fun

If you still have time, try to trace other things and/or improve the instrumentation. For example:

• Maybe we would like to have an overarching span when calling the status operation (the one polling the status of the order) at the StatusServlet and the KitchenConsumer.getDonuts call, like we did in the previous step with OrderServlet and KitchenConsumer.addDonut
• The automatic span names are sometimes overly general (e.g., "post"): try to override them with something more revealing
• Add span tags to make traces more self-descriptive and contextualized

Thanks for playing, and welcome to OpenTracing!

Thanks for joining us in this walkthrough! Hope you enjoyed it. If you did, let us know, and consider spreading the love!

A great way to get the feel for OpenTracing is to try your hand at instrumenting the OSS servers, frameworks, and client libraries that we all share. If you make one, consider adding it to the growing ecosystem at http://github.com/opentracing-contrib. If you maintain a library yourself, plase consider adding built-in OT support.

We also need walkthroughs for languages other than Golang. Feel free to reuse the client, protobufs, and other assets from here if you'd like to make one.

For a more detailed explanation of OSS Instrumentation, check out the Turnkey Tracing proposal at http://bit.ly/turnkey-tracing.

Aloha!