Issue with passing trace_entity when using ThreadPoolExecutor in lambda

I followed the instructions in the readme for passing the trace entity in a ThreadPoolExecutor but this does not appear to work properly in a lambda. The first time a thread does some work the trace_entity does not stick and the subsegment gets put under the root node not the one it should be assiciated with. The next time a thread picks up a job everything seems to operate fine.

I dug into the code a bit and it seems like there is a custome context class for lambdas but it does not override `set_trace_entity` to work properly with the other changes in the class. To workaround the issue I have found calling `get_trace_entity` before setting sets the internal state up properly.

example workaround from readme
```
def load_url(url, trace_entity):
xray_recorder.get_trace_entity() # Workaround to setup internal state
xray_recorder.set_trace_entity(trace_entity)

resp = requests.get(url)

xray_recorder.clear_trace_entities()
return resp
```

Beginne mit dem ThreadPoolExecutor-Beispiel in der README und untersuche die für Lambdas verwendete benutzerdefinierte Kontextklasse, insbesondere wie Trace-Entities initialisiert und gesetzt werden. Reproduziere den ersten von einem Thread bearbeiteten Job und verifiziere, dass sein Subsegment an die bereitgestellte Trace-Entity angehängt wird, ohne dass zuvor ein get_trace_entity-Aufruf erforderlich ist.

Vom Indexierungsmodell aus dem Issue-Text verfasst.