without this change we're explicitly requiring our users to take both web and actuator into their classpath. That makes little sense since not every app is a web app.
with this change we're making those deps optional
fixes#524
without this change when someone has a missing ErrorController bean then TraceHandlerInterceptor blows up
with this change that problem gets fixed
fixes#512
without this change there's a problem with the time when the SS is set on a span. Currently it's done in TraceFilter's finally block. The problem is that this code is executed after the response has been sent back to the client. Thus CR sometimes was set faster than SS (it doesn't make any sense from the logical point of view).
with this change we're introducing wrappers over the HttpServletResponse where we annotate the span with SS just after the response gets sent to the recipient.
fixes#492#431
without this change you couldn't register your own client (for example to set your ssl config) for Sleuth to reuse it
with this change we're reusing a registered feign client bean or creating a default one if necessary
fixes#502
* Add a new log method that allows for setting the time
* Clarify with time unit
* Simple test and spelling correction
* Consolidate "logic" and add a test for the other log method
Most changes are not very interesting to sleuth, though the server is
quite a lot better with UI fixes galore. For sleuth mainly, small api
unaffecting glitch fixes around the Span.Builder.
without this change it's pretty much impossible to enforce sampling for certain traces
with this change setting the X-B3-Flags to 1 for HTTP messages / spanFlags to 1 for messaging will override any sampling decisions
fixes#496
Local Ip determined from spring.commons InetUtils.
To enable local endpoint resolution via service discovery, DiscoveryClient must be present and 'spring.zipkin.service.locator.discovery' should be set to true.
FallbackHavingEndpointLocator was caching and recreating EndpointLocator in the same time - removed the caching part.
Fixes gh-403
without this tag it's impossible to discern from which server was the given span originated
with this change we're adding a tag in which we're passing the instance id value. The value can be taken either from Cloud Foundry or from a concatanation of some local properties like instance_id / application name / application port etc.
fixes#369
without this change the non web apps can't start
with this change the missing bean gets registered
also in terms of non boot apps if there is no port or address set we're providing some default values. In terms of the service name one can always use the override via the spring.zipkin.service.name property
fixes#32
without this only either service discovery service id / spring.application.name can be chosen as a service name for zipkin
with this change you can pass spring.zipkin.service.name property to change override that both for HTTP and Stream collectors
fixes#324
without this change during asynchronous communication some components are trying to detach or close spans that were aready detached. This leads to exceptions utils warnings and spans were not closed.
with this change we're adding additional checks to ensure that we're tracing
fixes#447
without this change Zipkin doesn't properly visualize the span on the server side.
with this change we're setting SA only if peer tag is set.
fixes#481
without this change we have 2 different properties. One is `spring.sleuth.metric` and the other is `spring.sleuth.metrics`.
with this change we're introducing conditionality on `spring.sleuth.metric.enabled`
fixes#477
without this change an explicit new span is created on the server side. Its name is equal to the method name of the controller. It introduces some nice advantages in terms of readability of trace.
with this change we're continuing a previous span on the server side. We're attaching the tags and logs to that span with information about controller class and controller name. Also events related to start and finish of the controller are there.
fixes#471#469#427
With this change we change the approach to continue spans (if they already exist) instead of starting new ones. The RPC spans would still be generated but Hystrix, Async, Callables / Runnables will reuse an existing span if there is one.
the scenarios are as follows:
Assuming that we have a trace X with span Y
* if you used tracer.wrap(Callable) or trace.wrap(Runnable) then:
* previously you'd get a span Z created when the Callable / Runnable is executed
* with this change you'll continue the span Y
* if you used a HystrixCommand then
* previously you'd get a span Z together with added tags when the command got executed
* with this change you'll continue the span Y and the tags will be added to span Y
* if you used a ExecutorService then
* previously you'd get a span Z together with added tags when a method from ExecutorService got executed
* with this change you'll continue the span Y and the tags will be added to span Y
Assuming that there was no span then everything will work as previously.
In order to create a new span you just have to create it manually. Example of creating a new span for an `@Async` annotated method.
```
// obviously you should inject via constructor ;)
@Autowired Tracer tracer;
@Async
public Future<String> foo() {
Span span = tracer.createSpan("newSpan");
try {
// do your stuff
} finally {
this.tracer.close(span);
}
}
```
fixes#174
