traceId / spanId generated are changed post WebClient call

We are using ReactorNettyHttpTracing with our Spring Web + Spring WebFlux + Spring Cloud Sleuth application. The traceId / spanId generated are getting changed post the WebClient call.

Spring Boot v2.6.1 and Spring Cloud v3.1.0

How do we retain the same traceId and spanId for the entire Spring Web request?

    @Bean
    public ReactorNettyHttpTracing reactorNettyHttpTracing(HttpTracing httpTracing) {
        return ReactorNettyHttpTracing.create(httpTracing);
    }

    WebClient.builder()
            .clientConnector(new ReactorClientHttpConnector(tracing.decorateHttpClient(httpClient)))
            .defaultHeader(HttpHeaders.CONTENT_TYPE, MediaType.APPLICATION_JSON_VALUE)
            .filters(exchangeFilters -> { 
                exchangeFilters.add(exchangeFilter.logRequest());
                exchangeFilters.add(exchangeFilter.logResponse());
            }).build();

[MohithMarisetti]

I’m having an almost similar issue with Spring cloud sleuth 3.x.x. I recently upgraded to Spring boot 2.6.6 and Sleuth 3.x.x. After the upgrade I see my HandlerInterceptor class is giving a different traceId in the afterCompletion() method. More info available here

@marcingrzejszczak @syedyusufh Any solution for this problem?

[marcingrzejszczak]

From what I see in this sample https://github.com/syedyusufh/sleuth-web-webflux.git it has nothing to do with WebClient. It’s more with the fact that the TraceValve didn’t properly support async dispatch. That means calling TraceValve twice with an span attribute already set once would not reuse span but create the new one.