一、该问题的重现步骤是什么？
1. bladex cloud版本，从bladex3.4.0升级到bladex4.5.0
2. 除了blade-admin模块之外，其他模块的/actuator/prometheus接口返回中缺失了jvm相关的指标，导致grafana的jvm仪表盘中没有数据
3. 为了排除干扰，从头搭建了bladex官方最新版纯净环境，效果一致，也是除了blade-admin模块之外都缺了jvm的指标
二、你期待的结果是什么？实际看到的又是什么？
我期待的结果是有jvm相关的一系列指标，比如blade-admin模块的返回节选：
HTTP/1.1 200 OK Content-Type: text/plain;version=0.0.4;charset=utf-8 Content-Length: 24963 Cache-Control: no-cache, no-store, max-age=0, must-revalidate Pragma: no-cache Expires: 0 X-Content-Type-Options: nosniff X-XSS-Protection: 0 Referrer-Policy: no-referrer connection: close # HELP jvm_threads_daemon_threads The current number of live daemon threads # TYPE jvm_threads_daemon_threads gauge jvm_threads_daemon_threads{application="blade-admin",} 122.0 # HELP jvm_compilation_time_ms_total The approximate accumulated elapsed time spent in compilation # TYPE jvm_compilation_time_ms_total counter jvm_compilation_time_ms_total{application="blade-admin",compiler="HotSpot 64-Bit Tiered Compilers",} 19907.0 # HELP executor_pool_core_threads The core number of threads for the pool # TYPE executor_pool_core_threads gauge executor_pool_core_threads{application="blade-admin",name="applicationTaskExecutor",} 8.0 # HELP executor_pool_max_threads The maximum allowed number of threads in the pool # TYPE executor_pool_max_threads gauge executor_pool_max_threads{application="blade-admin",name="applicationTaskExecutor",} 2.147483647E9 # HELP jvm_memory_used_bytes The amount of used memory # TYPE jvm_memory_used_bytes gauge jvm_memory_used_bytes{application="blade-admin",area="nonheap",id="CodeHeap 'profiled nmethods'",} 2.5495296E7 jvm_memory_used_bytes{application="blade-admin",area="heap",id="G1 Survivor Space",} 5606000.0 jvm_memory_used_bytes{application="blade-admin",area="heap",id="G1 Old Gen",} 3.5953152E7 jvm_memory_used_bytes{application="blade-admin",area="nonheap",id="CodeHeap 'non-nmethods'",} 1376256.0 jvm_memory_used_bytes{application="blade-admin",area="nonheap",id="Metaspace",} 7.1174928E7 jvm_memory_used_bytes{application="blade-admin",area="heap",id="G1 Eden Space",} 0.0 jvm_memory_used_bytes{application="blade-admin",area="nonheap",id="CodeHeap 'non-profiled nmethods'",} 6805248.0 jvm_memory_used_bytes{application="blade-admin",area="nonheap",id="Compressed Class Space",} 9008928.0 # HELP jvm_gc_pause_seconds Time spent in GC pause # TYPE jvm_gc_pause_seconds summary jvm_gc_pause_seconds_count{action="end of minor GC",application="blade-admin",cause="Metadata GC Threshold",gc="G1 Young Generation",} 1.0 jvm_gc_pause_seconds_sum{action="end of minor GC",application="blade-admin",cause="Metadata GC Threshold",gc="G1 Young Generation",} 0.006 jvm_gc_pause_seconds_count{action="end of minor GC",application="blade-admin",cause="G1 Evacuation Pause",gc="G1 Young Generation",} 2.0 jvm_gc_pause_seconds_sum{action="end of minor GC",application="blade-admin",cause="G1 Evacuation Pause",gc="G1 Young Generation",} 0.031 # HELP jvm_gc_pause_seconds_max Time spent in GC pause # TYPE jvm_gc_pause_seconds_max gauge jvm_gc_pause_seconds_max{action="end of minor GC",application="blade-admin",cause="Metadata GC Threshold",gc="G1 Young Generation",} 0.0 jvm_gc_pause_seconds_max{action="end of minor GC",application="blade-admin",cause="G1 Evacuation Pause",gc="G1 Young Generation",} 0.009 # HELP spring_security_filterchains_active_seconds_max # TYPE spring_security_filterchains_active_seconds_max gauge spring_security_filterchains_active_seconds_max{application="blade-admin",spring_security_filterchain_position="0",spring_security_filterchain_size="0",spring_security_reached_filter_name="none",spring_security_reached_filter_section="before",} 0.0 spring_security_filterchains_active_seconds_max{application="blade-admin",spring_security_filterchain_position="0",spring_security_filterchain_size="0",spring_security_reached_filter_name="none",spring_security_reached_filter_section="after",} 0.0 # HELP spring_security_filterchains_active_seconds # TYPE spring_security_filterchains_active_seconds summary spring_security_filterchains_active_seconds_active_count{application="blade-admin",spring_security_filterchain_position="0",spring_security_filterchain_size="0",spring_security_reached_filter_name="none",spring_security_reached_filter_section="before",} 0.0 spring_security_filterchains_active_seconds_duration_sum{application="blade-admin",spring_security_filterchain_position="0",spring_security_filterchain_size="0",spring_security_reached_filter_name="none",spring_security_reached_filter_section="before",} 0.0 spring_security_filterchains_active_seconds_active_count{application="blade-admin",spring_security_filterchain_position="0",spring_security_filterchain_size="0",spring_security_reached_filter_name="none",spring_security_reached_filter_section="after",} 0.0 spring_security_filterchains_active_seconds_duration_sum{application="blade-admin",spring_security_filterchain_position="0",spring_security_filterchain_size="0",spring_security_reached_filter_name="none",spring_security_reached_filter_section="after",} 0.0
目前看到的完整结果如下
HTTP/1.1 200 OKConnection: close Content-Type: text/plain;version=0.0.4;charset=utf-8 Content-Length: 6547 Date: Thu, 15 May 2025 08:39:07 GMT # HELP undertow_connectors_bytes_sent_bytes # TYPE undertow_connectors_bytes_sent_bytes gauge undertow_connectors_bytes_sent_bytes{protocol="http",} 0.0 # HELP executor_pool_core_threads The core number of threads for the pool # TYPE executor_pool_core_threads gauge executor_pool_core_threads{name="applicationTaskExecutor",} 8.0 executor_pool_core_threads{name="taskScheduler",} 1.0 # HELP executor_queued_tasks The approximate number of tasks that are queued for execution # TYPE executor_queued_tasks gauge executor_queued_tasks{name="applicationTaskExecutor",} 0.0 executor_queued_tasks{name="taskScheduler",} 0.0 # HELP undertow_connectors_processing_time_max_ms # TYPE undertow_connectors_processing_time_max_ms gauge undertow_connectors_processing_time_max_ms{protocol="http",} 0.0
-
给业务服务加一个这个bean配置可开启jvm指标
@Configuration public class SysConfig { @Bean public InitializingBean forcePrometheusPostProcessor(BeanPostProcessor meterRegistryPostProcessor, PrometheusMeterRegistry registry) { return () -> meterRegistryPostProcessor.postProcessAfterInitialization(registry, ""); } }
讨论(1)
-
我自己也找了一些信息，问题的原因是某些依赖的模块使用MeterRegistry的方式，导致MeterRegistry被过早初始化，此时MeterBinder还没有初始化完，错过了绑定的时机。解决方法就是在spring框架生命周期比较靠后的时机重新触发一次绑定，或者把要用的MeterBinder手工绑定一下。
使用ChatGPT对比了InitializingBean的方式和手工绑定的方式，ChatGPT更建议手工绑定的方式，理由是不破坏spring的生命周期，两种方式都能解决问题，手工绑定如下：
@Bean public ApplicationRunner bindJvmMetrics(MeterRegistry registry) { return args -> { try { // 手工new需要的MeterBinder并绑定 new JvmThreadMetrics().bindTo(registry); new JvmCompilationMetrics().bindTo(registry); new JvmMemoryMetrics().bindTo(registry); new JvmGcMetrics().bindTo(registry); new JvmHeapPressureMetrics().bindTo(registry); new JvmInfoMetrics().bindTo(registry); new ProcessorMetrics().bindTo(registry); new ClassLoaderMetrics().bindTo(registry); new UptimeMetrics().bindTo(registry); new FileDescriptorMetrics().bindTo(registry); new LogbackMetrics().bindTo(registry); } catch (Exception e) { log.error(e.getMessage(), e); } }; }
讨论(0)
加载中...
苏公网安备 32041102000998号