Spring AI The SimpleLoggerAdvisor and the custom advisor have failed.

Bug description Firstly, when I was using the custom Advisor, I found that the debugging did not take effect. Then, in order to verify whether it was a problem with my code, I used the SimpleLoggerAdvisor from the dependencies for debugging, but still found that it did not work.

SimpleLoggerAdvisor Configuration Example

@Bean
public ChatClient zhipuAiChatClient(@Qualifier("zhiPuAiChatModel") ChatModel chatModel, ToolCallback[] toolCallbacks) {
    ChatClient chatClient = ChatClient.builder(chatModel)

            // 实现 Logger 的 Advisor
            .defaultAdvisors(
                    new SimpleLoggerAdvisor()
            )
            .defaultToolCallbacks(toolCallbacks)
            .build();
    return chatClient;
}

Custom Adviso Configuration Example

@Bean
public ChatClient zhipuAiChatClient(@Qualifier("zhiPuAiChatModel") ChatModel chatModel, ToolCallback[] toolCallbacks) {
    ChatClient chatClient = ChatClient.builder(chatModel)

            // 实现 Logger 的 Advisor
            .defaultAdvisors(
                    new MyLogAdvisor()
            )
            .defaultToolCallbacks(toolCallbacks)
            .build();
    return chatClient;
}

Custom Adviso Code Example

/**
 * @author mi
 * @data 2025/6/12 9:07
 * @version 1.0
 */
@Slf4j
public class MyLogAdvisor implements CallAdvisor, StreamAdvisor {

    private void before(ChatClientRequest chatClientRequest) {
        log.info("请求prompt：\n{}", chatClientRequest.prompt().getContents());
    }

    @Override
    public ChatClientResponse adviseCall(ChatClientRequest chatClientRequest, CallAdvisorChain callAdvisorChain) {
        this.before(chatClientRequest);
        ChatClientResponse chatClientResponse = callAdvisorChain.nextCall(chatClientRequest);

        if (chatClientResponse.chatResponse() != null) {
            String responseText = chatClientResponse.chatResponse().getResult().getOutput().getText();
            log.info("AI响应：\n{}", responseText);
        } else {
            log.error("AI响应失败：\n{}", JSONUtil.toJsonStr(chatClientResponse));
        }

        return chatClientResponse;
    }

    @Override
    public Flux<ChatClientResponse> adviseStream(ChatClientRequest chatClientRequest, StreamAdvisorChain streamAdvisorChain) {
        this.before(chatClientRequest);

        Flux<ChatClientResponse> chatClientResponseFlux = streamAdvisorChain.nextStream(chatClientRequest);

        Mono<List<ChatClientResponse>> collectedFlux = chatClientResponseFlux.collectList();
        collectedFlux
                .doOnNext(responses -> {
                    StringBuilder resContent = new StringBuilder();
                    responses.forEach(resContent::append);
                    log.info("AI响应：\n{}", resContent);
                })
                .subscribe();
        return chatClientResponseFlux;
    }

    @Override
    public String getName() {
        return "MyLogAdvisor";
    }

    @Override
    public int getOrder() {
        return Integer.MAX_VALUE;
    }
}

Environment spring.ai.version 1.0.0 jdk 17

Minimal Complete Reproducible example

chatClient.prompt("hello world").advisors(new SimpleLoggerAdvisor()).stream().content();

Comment From: sunyuhan1998

If you wish to use SimpleLoggerAdvisor, you will need to adjust the project's logging level to debug.