flink监控指标大屏搭建

Flink Metrics

         Flink12

metrics.reporters: prom
metrics.reporter.prom.class: org.apache.flink.metrics.prometheus.PrometheusReporter
metrics.reporter.prom.port: 9213-9214

 

Flink17配置

      和flink12的配置方式不同

jobmanager.rpc.address: flink-node1

jobmanager.rpc.port: 6123


jobmanager.memory.process.size: 4096m
jobmanager.memory.jvm-metaspace.size: 2048m


taskmanager.memory.task.heap.size: 20480m
taskmanager.memory.managed.size: 1024m
taskmanager.memory.framework.off-heap.size: 2048m
taskmanager.memory.jvm-metaspace.size: 2048m
taskmanager.numberOfTaskSlots: 300


parallelism.default: 1

state.checkpoints.num-retained: 2


jobmanager.execution.failover-strategy: region

rest.bind-address: 0.0.0.0

io.tmp.dirs: /home/taishi/nfsdata/flink/flink-temp
historyserver.web.tmpdir: /home/taishi/nfsdata/flink/flink-web-tmp

jobmanager.archive.fs.dir: file:///home/taishi/nfsdata/flink/flink-history

historyserver.web.port: 8882
historyserver.archive.fs.dir: file:///home/taishi/nfsdata/flink/flink-history

env.java.home: /usr/lib/java/jdk1.8.0_191
web.upload.dir: /home/taishi/nfsdata/flink/flink-web-jar

metrics.reporters: prom
#metrics.reporter.prom.class: org.apache.flink.metrics.prometheus.PrometheusReporter
#metrics.reporter.prom.factory.class: org.apache.flink.metrics.prometheus.PrometheusReporter
metrics.reporter.prom.factory.class: org.apache.flink.metrics.prometheus.PrometheusReporterFactory
metrics.reporter.prom.port: 9213-9214
classloader.check-leaked-classloader: false


restart-strategy.fixed-delay.attempts: 3
restart-strategy.fixed-delay.delay: 10s
restart-strategy: fixed-delay

cluster.evenly-spread-out-slots: true
env.java.opts: -Dlog4j2.formatMsgNoLookups=true

    

    

调试方式

       查看flink启动的详细日志

           

           

           

           

 展示监控大屏异常

          

          

          

           curl -X POST -g 'http://localhost:9090/api/v1/admin/tsdb/delete_series?match[]={exported_job="tf_flinkmonitor09cd6b3d26e1751eaa32dc1cb4371ec3"}'

          

          

      正常展示大屏数据

      

      1.删除多余的job

      2.增加指标查询过滤条件