Tracing improvements

[pavolloffay]

Hi,

I went through the code and I would like to propose a couple of improvements in the tracing:

• Use instrumentation to automatically propagate the headers. When a service does a downstream call the context (trace ids) have to be propagated. The guide https://github.com/dwojciec/debugging-lab-summit-2019/blob/master/instructions/jaeger.md#enabling-distributed-context-propagation indicates to propagate the headers manually, this can be fully automated by using instrumentation library from https://github.com/opentracing-contrib and a tracer which supports B3 propagation - Jaeger java client with JAEGER_PROPAGATION=b3. This step can showcase how instrumentation helps to propagate the context.
• Once service is instrumented it will produce more tracing data from inside the service. For example if there is an error it will be possible to see the error message or redirects within the execution in the service.
• Add business data to a span. In the controller it will be possible to inject a tracer and add for instance customer ID into the span which can be used as a query in Jaeger UI data.tracer.activeSpan().setTag("customer-id", customerId).
• Smart istio routing using OpenTracing baggage https://redhat-developer-demos.github.io/istio-tutorial/istio-tutorial/1.1.x/4advanced-routerules.html#canarydeploymentuseragent

Tracing Istio tutorial: https://redhat-developer-demos.github.io/istio-tutorial/istio-tutorial/1.1.x/3monitoring-tracing.html#tracing