Radoslaw Busz created SPARK-37147: ------------------------------------- Summary: MetricsReporter producing NullPointerException when element 'triggerExecution' not present in Map[] Key: SPARK-37147 URL: https://issues.apache.org/jira/browse/SPARK-37147 Project: Spark Issue Type: Bug Components: Structured Streaming Affects Versions: 3.2.0 Reporter: Radoslaw Busz
The exception occurs in MetricsReporter when it tries to register gauges using lastProgress of each stream. This problem was partially fixed in https://issues.apache.org/jira/browse/SPARK-22975 but in introduced a NPE when 'triggerExecution' element is not present in durationMS map: {code:java} registerGauge("latency", _.durationMs.get("triggerExecution").longValue(), 0L) {code} I find it difficult to reproduce this every time but it happens every few restarts when structured streaming uses a slow event source (very little or no events). In my case it breaks metric reporting via Codehale/Dropwizard and generates multiple stacktraces such as: {code:java} 21/09/16 09:51:36 ERROR ScheduledReporter: Exception thrown from GangliaReporter#report. Exception was suppressed. java.lang.NullPointerException at org.apache.spark.sql.execution.streaming.MetricsReporter.$anonfun$new$3(MetricsReporter.scala:43) at org.apache.spark.sql.execution.streaming.MetricsReporter.$anonfun$new$3$adapted(MetricsReporter.scala:43) at scala.Option.map(Option.scala:230) at org.apache.spark.sql.execution.streaming.MetricsReporter$$anon$1.getValue(MetricsReporter.scala:68) at com.codahale.metrics.ganglia.GangliaReporter.reportGauge(GangliaReporter.java:353) at com.codahale.metrics.ganglia.GangliaReporter.report(GangliaReporter.java:240) at com.codahale.metrics.ScheduledReporter.report(ScheduledReporter.java:237) at com.codahale.metrics.ScheduledReporter.lambda$start$0(ScheduledReporter.java:177) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) {code} I'm happy to implement a fix. -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org