The tracing systests have been very unstable for me, failing more often then
not with failures in
org.apache.cxf.systest.jaxrs.tracing.brave.BraveTracingTest. The test it
eventually fails in in that class seems relatively random. In each case, the
number of spans is greater than what is expected. Is anyone else seeing that?
I tried digging into it and it LOOKS like the calls to "get
/bookstore/books/async” are leaving an “inFlight” span in the Tracer. That
span is then delivered at some point in the future which then causes a test to
fail.
0:
{"traceId":"3a5f1a7d2de45f49","parentId":"3a5f1a7d2de45f49","id":"b0f4e2ddef4251f5","name":"processing
books","timestamp":1535566440433652,"duration":200595,"localEndpoint":{"serviceName":"unknown","ipv4":"192.168.1.180"}}
1:
{"traceId":"3a5f1a7d2de45f49","id":"3a5f1a7d2de45f49","kind":"SERVER","name":"get
/bookstore/books/async","timestamp":1535566440423025,"duration":212695,"localEndpoint":{"serviceName":"unknown","ipv4":"192.168.1.180"},"tags":{"http.method":"GET","http.path":"/bookstore/books/async"}}
Tracer{inFlight=[{"traceId":"3a5f1a7d2de45f49","id":"3a5f1a7d2de45f49","localEndpoint":{"serviceName":"unknown","ipv4":"192.168.1.180"},"shared":true}],
reporter=org.apache.cxf.systest.brave.TestSpanReporter@1da2cb77}
Is there something missing on the sever side in the async case to close off the
span or something?
--
Daniel Kulp
[email protected] <mailto:[email protected]> - http://dankulp.com/blog
<http://dankulp.com/blog>
Talend Community Coder - http://talend.com <http://coders.talend.com/>
