We're currently seeing our services spend a large amount of time inside of the lightstep sdk
serializing reports when they instrument themselves with tracing.  This is in large part due to
the fact that lightstep is using the pure python implemenation of Thrift.

Fixes lightstep#89

When using THttpClient.THttpClient as the transport, an invalid status code could cause memory usage to spike. This behaviour is corrected via TTransport.TBufferedTransport, so this check is no longer needed.

Signed-off-by: Alex Boten <[email protected]>

…hrift_protocol

Signed-off-by: Alex Boten <[email protected]>

…hrift_protocol