Fix blocking code (jaegertracing#191)

* fixed blocking code in Reporters/RemoteReporter.cs

Signed-off-by: Bojan Pantovic <bojan.pantovic@redbox.com>

* replace blocking collection with BufferBlock to make everything async

Signed-off-by: Bojan Pantovic <bojan.pantovic@redbox.com>

* code review changes

Signed-off-by: Bojan Pantovic <bojan.pantovic@redbox.com>

* code review changes jaegertracing#2

Signed-off-by: Bojan Pantovic <bojan.pantovic@redbox.com>

* Fix NuGET related build warnings

Signed-off-by: Kraemer, Benjamin <falco20019@hotmail.com>
Signed-off-by: Bojan Pantovic <bojan.pantovic@redbox.com>

Co-authored-by: bojan.pantovic@redbox.com <bojan.pantovic@redbox.com>
Co-authored-by: Kraemer, Benjamin <falco20019@hotmail.com>

Fix Lost Spans on Close (jaegertracing#182)

* Make InMemorySender Asynchronous

To demonstrate lost span problem (to be shortly reported). When AppendAsync is
actually asynchronous, some spans are lost on Tracer Dispose. This change should
have no effect, but causes RemoteReporterTests.TestRemoteReporterFlushesOnClose
to fail intermittently.

Signed-off-by: phxnsharp <nsharp@phoenix-int.com>

* Do not use Task.Factory.StartNew

For some reason awaiting the task returned always completes immediately.

Signed-off-by: phxnsharp <nsharp@phoenix-int.com>

* Add Thread Safety to InMemorySender

This didn't help reliability, but I believe it is necessary because
different threads may be adding and flushing at the same time.

Signed-off-by: phxnsharp <nsharp@phoenix-int.com>

* Style cleanup

Use slightly better style for _blocker. Also, turns out async in
Flush causes other issues, removed for now.

Signed-off-by: phxnsharp <nsharp@phoenix-int.com>

* Fix Intermittent Test Failure

Found several problems that were contributing:
 - RemoteControlledSampler was using a long lived Wait() call on a thread pool thread, which was consuming Thread Pool threads and causing delays as the thread pool was exausted
 - RemoteReporter.ProcessQueueLoop does have long lived waits due to the BlockingQueue. Reworked it to work correctly on a LongLived thread without consuming thread pool threads.
 - Many of the tests were leaving behind the ProcessQueueLoop because the blocker was never released.

There is still an intermittent test failure, but that was there before I started changing code.

Signed-off-by: phxnsharp <nsharp@phoenix-int.com>
Co-authored-by: Benjamin Krämer <falco20019@hotmail.com>

Improve HttpSender flexibility (jaegertracing#184)

- WithHttpHandler
- WithCertificates
- WithUserAgent

Signed-off-by: Kraemer, Benjamin <falco20019@hotmail.com>

Replace JAEGER_SAMPLER_MANAGER_HOST_PORT with JAEGER_SAMPLING_ENDPOINT (

jaegertracing#165)

* Replace JAEGER_SAMPLER_MANAGER_HOST_PORT with JAEGER_SAMPLING_ENDPOINT
* Readded support for deprecated JAEGER_SAMPLER_MANAGER_HOST_PORT
* Adjusted tests for real-world sampler results

Add icon url for Jaeger nuget package (jaegertracing#160)

Signed-off-by: Philipp Bocharov <bocharovf@gmail.com>

Set default SamplingManager to HttpSamplingManager (jaegertracing#157)

Signed-off-by: Rustam Valiullin <regeldso@users.noreply.github.com>

Ignore Jaeger internal traces in HttpSender (jaegertracing#154) (jaeg…

…ertracing#156)

Ignore Jaeger internal traces in HttpSender.

Mark Jaeger client internal requests made by HttpSender with
"ot-ignore" property which prevents requests from being instrumented by
OpenTracing.Contrib.NetCore package and hides them in UI.

Signed-off-by: Philipp Bocharov <bocharovf@gmail.com>

Extending the range of supported versions of newtonsoft (jaegertracin…

…g#150)

Now depended on Newtonsoft.Json 9.0.1 and above

Signed-off-by: SanSYS <i.sansys@gmail.com>

Added support for 128bit to B3 decoder (jaegertracing#144)

Using FromIConfiguration throughout complete cascade (jaegertracing#138)